| hggdh | heh. Either sense would work | 00:00 |
|---|---|---|
| hggdh | Daviey, see byobu screen 2 | 00:02 |
| hggdh | Daviey, I deregistered cluster1 (just in case) and tried again to publish. Went splat as usual | 00:03 |
| RoAkSoAx | who can post in Ubuntu Server Blog? WOuldn't it be nice to post about PowerNap too? :) | 00:06 |
| Daviey | hmm | 00:06 |
| Daviey | hggdh, what was 10.55.55.8? | 00:07 |
| hggdh | Daviey, this was (I expect) your SC | 00:08 |
| Daviey | RoAkSoAx, It's not entirely clear... Technically it should be someone from ~ubuntu-members, as i don't think we had an ack to add it | 00:08 |
| Daviey | hggdh, interesting.... | 00:08 |
| Daviey | 19:06:16 DEBUG [ChannelUtil:New I/O client boss #2] java.net.NoRouteToHostException: No route to host | 00:08 |
| Daviey | java.net.NoRouteToHostException: No route to host | 00:08 |
| Daviey | (i just rebooted it) | 00:08 |
| Daviey | So euca *IS* trying to play with it | 00:08 |
| hggdh | even after I deregistered the beast? | 00:09 |
| Daviey | seems so | 00:09 |
| hggdh | OK. Is there anything kosher? The more we poke the worst it seems | 00:09 |
| twb | Water probably is | 00:10 |
| twb | Unless a pig died in it or something, I guess | 00:10 |
| Daviey | hggdh, I think the whole rack should be flushed tbh | 00:10 |
| Daviey | hggdh, at the moment, i don't care if multiple clouds on the same subnet causes it to go bang... :) | 00:11 |
| Daviey | hggdh, It's late for me, and i need to step away from the computer before my eyes explode. | 00:11 |
| Daviey | If you can't get me before you go, can you send me an email with a sitrep? | 00:11 |
| hggdh | Daviey, certainly | 00:15 |
| hggdh | I am going to restart from scratch | 00:15 |
| * Patrickdk watchs everyones heads explode | 00:15 | |
| * hggdh was doing good until getting to test euca... | 00:17 | |
| Patrickdk | just working on the iscsi tests | 00:18 |
| Patrickdk | attempting to figure out why the fail | 00:18 |
| * hggdh grabs some popcorn and sits down to watch Patrickdk's head go the same route and Daviey's and self's | 00:27 | |
| Daviey | :( | 00:27 |
| Patrickdk | heh, I have been able to solve many issues, but dhcp/pxe/iscsi root, is pretty complex, and not much room for diagnosing | 00:28 |
| hggdh | Daviey, bug 728145 opened | 00:31 |
| uvirtbot | Launchpad bug 728145 in eucalyptus "euca tries to register a SC, fails, but adds it in" [Undecided,New] https://launchpad.net/bugs/728145 | 00:31 |
| hggdh | now for a Brand New ISO Install... | 00:31 |
| Daviey | hggdh, make sure you kill the two machines i was using | 00:31 |
| hggdh | Daviey, hostnames, please | 00:32 |
| SpamapS | Daviey: shouldn't you be in bed? :) | 00:33 |
| hggdh | Daviey, never mind, I will shut them down, PXE is set to default on all except the two I am installing | 00:33 |
| hggdh | SpamapS, a long time ago. I am always amazed on how he survives | 00:34 |
| SpamapS | Probably safer to stay on the computer rathern than have to face the wrath of the wife who has been waiting all night for him. ;) | 00:35 |
| SpamapS | I know I get that one. ;) | 00:35 |
| Daviey | SpamapS, yeah... | 00:35 |
| === TuxIce is now known as AdamDV-iPhone2 | ||
| Daviey | hggdh, mabolo and marula | 00:36 |
| Daviey | SpamapS, see comment on bug #711425 | 00:36 |
| uvirtbot | Launchpad bug 711425 in sysvinit "portmap does not stop during shutdown, causing possible root fs corruption" [High,In progress] https://launchpad.net/bugs/711425 | 00:36 |
| hggdh | SpamapS, it is just shy of 1900 here, and my wife is already unhappy... | 00:38 |
| SpamapS | Daviey: I'd be willing to bet sysvinit is FTBFS then | 00:38 |
| Daviey | yeah :) | 00:38 |
| Daviey | SpamapS, did you test build before submitting the debdiff? | 00:39 |
| SpamapS | Daviey: probably not given that I only changed shell scripts | 00:39 |
| Daviey | fair enough :) | 00:39 |
| Patrickdk | yay, I got login :) | 00:40 |
| SpamapS | Daviey: might be something that was dropped in 2.6.38 | 00:41 |
| twb | Last time I looked, portmap was done in upstart as at 10.04... | 00:47 |
| SpamapS | portmap is an upstart job yes | 00:47 |
| SpamapS | but shutdown is not in upstart :( | 00:47 |
| SpamapS | so sysvinit has to stop portmap | 00:47 |
| twb | Er, don't you mean "portmap should have a `stop on' line"? | 00:48 |
| SpamapS | it has that now | 00:48 |
| twb | Hum, OK | 00:48 |
| Patrickdk | one little fix, corrects 4 failed, but I doubt the *little* fix is proper :) | 00:48 |
| SpamapS | stop on unmounted-remote-filesystems | 00:48 |
| Patrickdk | #728088 | 00:48 |
| twb | I'm *so* sick of upstart, man | 00:49 |
| SpamapS | twb: the issue is that portmap must not be killed until all NFS filesystems are unmounted | 00:49 |
| SpamapS | twb: its not so much upstart as the half-completed migration to it that is frustrating | 00:49 |
| twb | Yes, EXACTLY | 00:49 |
| SpamapS | twb: basically the migration isn't really done until everything that isn't a traditional network service is upstartified. | 00:50 |
| twb | After nearly four years and stuff still isn't migrated properly | 00:50 |
| twb | Hell, halt(8) is called from rc0.d still | 00:50 |
| SpamapS | I think there's been a lot of resistance and moaning about it but its not going anywhere so its time for us to just get'r'done | 00:51 |
| Patrickdk | heh, you should see the postfix vs upstart thread on that maillist :) | 00:51 |
| SpamapS | twb: Its likely that will be a topic of discussion for UDS-O | 00:51 |
| Patrickdk | they are cursing upstart cause it must monitor something to tell it's active | 00:52 |
| SpamapS | Patrickdk: I'm trying hard to stay out of that. | 00:52 |
| twb | SpamapS: surely 12.04 will be running systemd? | 00:52 |
| SpamapS | twb: that would be suicidal | 00:52 |
| SpamapS | twb: maybe 12.10 and later.. | 00:52 |
| SpamapS | twb: but dropping systemd in now would be WORSE than when upstart was dropped in during Karmic | 00:52 |
| twb | It'll be funny if Debian 7 releases with a complete systemd migration and 12.04 is still half-upstart, half-sysvinit | 00:53 |
| SpamapS | I do hope that we have upstart 2 w/ the state based interface so we can drop the wait for state hacks before 12.04 ... | 00:53 |
| SpamapS | twb: thats the point.. 11.10 should have a major push to complete the migration to upstart. | 00:54 |
| twb | Has remnant said whether he'll continue developing upstart at google? | 00:54 |
| SpamapS | He's just dropped upstart 1.0 yesterday | 00:54 |
| SpamapS | as in.. published | 00:54 |
| twb | That's not necesarily a "yes" :-) | 00:54 |
| SpamapS | and he did that as a sort of hand off to us so he can focus on 2.0 | 00:54 |
| SpamapS | Upstart seems to be a key part of what he's doing there. | 00:55 |
| SpamapS | but, he can neither confirm nor deny that | 00:55 |
| twb | Flipping corporations :-/ | 00:56 |
| SpamapS | well either way... its arguably highly reckless to just go switching inits *now* | 00:57 |
| Patrickdk | no one doing the raid1 tests? | 00:58 |
| SpamapS | better the devil we know. if upstart 2 scks or never lands.. so be it. | 00:58 |
| SpamapS | Patrickdk: I'd do them but my bandwidth is complete crap :-/ | 00:58 |
| SpamapS | hard enough just upgrading to the latests python 2.7 :-/ | 00:59 |
| SpamapS | twb: I have to agree that I'm tired of mopping up the upstart mess tho | 00:59 |
| twb | Especially when I see pere's insserv work in Debian | 01:00 |
| SpamapS | but I do see a clear path to it being finished | 01:00 |
| twb | Since that made the boot time significantly less, and didn't impact me as a user at all | 01:00 |
| hggdh | Daviey, you can go to bed... uec-publish-tarball now works :-) | 01:02 |
| patdk-lap | I'll try, but dunno if I have any time | 01:02 |
| Daviey | hggdh, pah | 01:03 |
| SpamapS | twb: I have to wonder how much of that has been influenced by upstart itself. | 01:03 |
| twb | I think it certainly precipitated interest, but I don't think the innserv methodology was in any way derived from upstart | 01:04 |
| twb | I mean, it looks at LSB headers and computes an ordering -- it's more like make than upstart | 01:05 |
| SpamapS | not derived, but certainly the existence of upstart and its perceived deficiencies has driven systemd development | 01:05 |
| hggdh | Patrickdk, I was going to do the raid1, but got stuck with euca | 01:05 |
| twb | Oh sorry, I thought you were talking about insserv, not systemd | 01:05 |
| SpamapS | I imagine it has also encouraged insserv work to improve things so that they have a compelling argument against upstart | 01:06 |
| hggdh | Patrickdk, if you can run them, I would be in your debt | 01:06 |
| * hggdh is getting to owe a lot of people around | 01:06 | |
| SpamapS | twb: I was making a giant leap.. loop closed now ;) | 01:06 |
| SpamapS | twb: the main point is just that any time an early leader falters, it encourages the new players who don't have to deal with all the mistakes of the first implementation | 01:09 |
| patdk-lap | hggdh, don't worry, I will collect :) | 01:09 |
| SpamapS | twb: upstart has found many of the really big potholes .. and systemd and insserv and others have been able to see it fall into those and can avoid them | 01:09 |
| twb | Like fortran | 01:09 |
| twb | Didn't insserv predate upstart? | 01:10 |
| SpamapS | this is funny.. may laptop on my lap is behind a bunch of plants.. if I'm doing a download and hold it high up in the air.. it speeds up from 512kbit to 2Mbit | 01:10 |
| twb | SpamapS: what would be funny is if you were using cat5 and that happened | 01:11 |
| SpamapS | s/funny/scary/ | 01:11 |
| SpamapS | insserv wasntreally going where upstart was... but yes.. it was leading, and it was passed | 01:12 |
| SpamapS | twb: I'm not exactly attached to upstart forever. I just think if it was going to be replaced in ubuntu, it needed to happen in 11.04 | 01:14 |
| twb | I totally admit that insserv's goals are conservative | 01:15 |
| twb | But at least it met them without giving me stress :-P | 01:15 |
| hggdh | RoAkSoAx, still there? | 01:19 |
| RoAkSoAx | hggdh: yep | 01:21 |
| RoAkSoAx | hggdh: what up :)? | 01:22 |
| RoAkSoAx | what's* | 01:22 |
| hggdh | RoAkSoAx, I am ready to test the change to powernap | 01:22 |
| RoAkSoAx | hggdh: you wanna do it now or tomorrow? | 01:22 |
| hggdh | RoAkSoAx, your call. I cannot test euca without it, though | 01:23 |
| RoAkSoAx | hggdh: I guess we can do it now then if you like | 01:23 |
| hggdh | RoAkSoAx, really, your call. I *know* it will fail no matter what, there is the DHCP issue still pending | 01:24 |
| hggdh | so we can do it tomorrow morning | 01:24 |
| RoAkSoAx | hggdh: ok then | 01:24 |
| hggdh | say around 0700 CST? | 01:24 |
| hggdh | 0800 EST | 01:25 |
| RoAkSoAx | hggdh: bettwe yet we will both be fresh | 01:25 |
| RoAkSoAx | hggdh: let';s make it 8.30 | 01:25 |
| RoAkSoAx | :) | 01:25 |
| hggdh | heh | 01:25 |
| RoAkSoAx | 7.30/8.30 | 01:25 |
| hggdh | my dogs wake up early, so I have no other option... | 01:25 |
| hggdh | RoAkSoAx, deal | 01:25 |
| RoAkSoAx | hggdh: alrighty!! have a good one them. I'm off now :P | 01:25 |
| hggdh | RoAkSoAx, thank you | 01:26 |
| RoAkSoAx | no worries :)_ | 01:26 |
| uvirtbot | New bug: #728167 in cloud-init (main) "failure on apt-get update should be warning" [Medium,Triaged] https://launchpad.net/bugs/728167 | 01:41 |
| emma_wall | Hi, im new to Linux! so i have a few silly question | 01:45 |
| patdk-lap | hmm, raid1 tests failing for me | 02:10 |
| patdk-lap | can't start raid in degraded mode | 02:10 |
| patdk-lap | hopefully I'm just doing it wrong | 02:10 |
| patdk-lap | mdadm: CREATE user root not found | 02:11 |
| patdk-lap | mdadm: CREATE group disk not found | 02:12 |
| SpamapS | lifeless: had any chance to try out the new cassandra packages? | 02:34 |
| SpamapS | patdk-lap: wha? | 02:34 |
| lifeless | SpamapS: yes, I commented in the bug | 02:34 |
| SpamapS | lifeless: oh good :) | 02:35 |
| lifeless | SpamapS: haproxy1.4 for lucid interest you? | 02:35 |
| SpamapS | lifeless: as in, a backport? | 02:36 |
| lifeless | yes | 02:37 |
| lifeless | or a CAT build | 02:37 |
| SpamapS | CAT build? | 02:38 |
| lifeless | ha proxy 1.3 cannot do close-mode | 02:38 |
| lifeless | canonical admin team | 02:38 |
| patdk-lap | spamaps, I probably did something wrong | 02:39 |
| patdk-lap | or, I hope I did :) | 02:39 |
| lifeless | SpamapS: without close mode, persistent connections show up as single sessions regardless of requests sent over them | 02:40 |
| lifeless | SpamapS: which breaks the concurrency limit | 02:40 |
| lifeless | because connection != request | 02:40 |
| lifeless | and haproxy explicitly says it doesn't see the requests beyond the first one (in 1.3) | 02:40 |
| lifeless | SpamapS: this is hindering our load normalisation for launchpad | 02:40 |
| patdk-lap | lifeless, adjust your keepalive settings? | 02:41 |
| patdk-lap | if you have keepalive off, only one request per connection is allowed | 02:42 |
| lifeless | I know this | 02:42 |
| lifeless | patdk-lap: sorry, I don't mean to be negative | 02:48 |
| lifeless | patdk-lap: its more complex than that - we have 4 tiers talking http | 02:48 |
| patdk-lap | na | 02:48 |
| patdk-lap | just wondering if it's something I'm missing :) | 02:48 |
| lifeless | apache (load balanced) to squid (load balanced) to haproxy (load balanced) to 80 appserver threads | 02:49 |
| patdk-lap | that seems like a mess | 02:49 |
| lifeless | apache handles the SSL | 02:49 |
| lifeless | squid does caching | 02:50 |
| lifeless | haproxy does load limiting | 02:50 |
| lifeless | but, 1.3 can't do it well enough to let us move to single threaded appservers. 1.4 can. | 02:50 |
| lifeless | if we reconfigure things in front of haproxy we may make it work for a bit, but if someone makes a mistake, we'll have a horrible failure mode | 02:51 |
| lifeless | I don't like horrible failure modes. | 02:51 |
| lifeless | patdk-lap: why do you say a mess? we serve a fair few hits daily | 03:10 |
| patdk-lap | I've never liked using apache proxy, or reverse squid cache | 03:15 |
| patdk-lap | I keep removing them to speed things up | 03:15 |
| lifeless | patdk-lap: do your sites run SSL ? | 03:16 |
| patdk-lap | yep | 03:16 |
| lifeless | how many application servers do you have? | 03:17 |
| patdk-lap | 2 | 03:17 |
| lifeless | what do you use to syncronise your SSL session cache | 03:18 |
| patdk-lap | nothing, I don't need to | 03:18 |
| SpamapS | lifeless: re the 1.4 build.. that makes sense. since haproxy is in main .. it should be an easy sell for backport to lucid | 03:18 |
| lifeless | then whenever a client switches application server, they will renegotiate. Thats up to 6 seconds from nz. | 03:18 |
| lifeless | SpamapS: how do I go about making it happen | 03:18 |
| patdk-lap | before they had some kind of apache squid apache thing, using 48 servers, 8 squid servers | 03:19 |
| patdk-lap | lifeless, they don't switch application servers, unless their ip changes | 03:19 |
| patdk-lap | since I have a limited amount of that, I just balance by ip instead of load | 03:19 |
| SpamapS | lifeless: realistically I'd be shocked if the package from maverick didn't build w/o changes on lucid | 03:19 |
| lifeless | patdk-lap: depends on the client, and if they have /any/ network interrupts. | 03:19 |
| SpamapS | lifeless: so an upload to the PPA that you guys use would probably be the only thing necessary. | 03:20 |
| patdk-lap | network interrupts shouldn't change their ip | 03:20 |
| lifeless | patdk-lap: you're pinning clients by the *client* ip ? | 03:20 |
| patdk-lap | sure | 03:20 |
| lifeless | patdk-lap: in which case, roaming clients - e.g. folk at UDS - will suffer regular SSL renegotiation. | 03:21 |
| lifeless | I mean, its your call how you do it. And I"m sure you solved problems you had. | 03:21 |
| lifeless | doesn't mean there aren't more issues present :) | 03:21 |
| patdk-lap | ya | 03:21 |
| patdk-lap | the biggest issue I found we has | 03:22 |
| patdk-lap | was duplication of caches between all the servers | 03:22 |
| patdk-lap | limiting it to less servers, fixed that hugely | 03:22 |
| lifeless | I would graph invalid ssl attempts - they will mostly be folk that have roamed trying wha they think is their existing session. | 03:22 |
| lifeless | patdk-lap: we'd need stupid-big caches to cache our data set | 03:25 |
| lifeless | patdk-lap: - its 300GB - | 03:25 |
| patdk-lap | nothing wrong with that | 03:25 |
| patdk-lap | but having the same stupid big cache on 40 servers | 03:25 |
| lifeless | if you had no locality of reference, yeah it would be a waste | 03:25 |
| lifeless | you'd get lots of cold-cache reads | 03:25 |
| patdk-lap | ya | 03:25 |
| patdk-lap | my issues was more, 500gig working set, 4gig ram, 48 servers | 03:26 |
| patdk-lap | basically everything was a cold read | 03:26 |
| lifeless | that would blow | 03:26 |
| patdk-lap | and wasting ram for squid | 03:27 |
| SpamapS | The working set that is entirely popular is quite rare. | 03:27 |
| lifeless | its mythical :) | 03:27 |
| patdk-lap | if there where no webcrawlers :) | 03:27 |
| lifeless | as flickr say, their peak time lasts all week :) | 03:28 |
| patdk-lap | hmm, this makes me wonder alittle | 03:28 |
| patdk-lap | cause I didn't think about it too much | 03:29 |
| patdk-lap | the 2 servers I replaced that whole mess with have 18gigs ram each in them | 03:29 |
| patdk-lap | but I run vm's on them, and the *webservers* only have 2gigs ram each | 03:29 |
| patdk-lap | doing more with less | 03:29 |
| patdk-lap | defently a cache issue | 03:29 |
| patdk-lap | less nfs hits iwth less servers hitting it up | 03:30 |
| SpamapS | patdk-lap: web crawlers are an easy case to handle. In a previous life we simply had our HTTP aware load balancer send all crawler user agents to a single server that we didn't mind going slow. ;) | 03:31 |
| SpamapS | patdk-lap: and we also heavily focused on making If-Modified-Since function properly. | 03:31 |
| patdk-lap | oh, I have never had an issue for my websites | 03:32 |
| patdk-lap | it's hosting other peoples sites | 03:32 |
| SpamapS | Never did convince anyone that we could make old pages hand out expires headers since they changed less frequently. :) | 03:32 |
| patdk-lap | where you shouldn't be modifing their stuff, to fix that :) | 03:32 |
| patdk-lap | hmm, I wonder how you would handle google these days though | 03:34 |
| patdk-lap | since they rate you based on crawl speed also | 03:34 |
| patdk-lap | or page load time | 03:34 |
| SpamapS | If their stuff is crap, you shouldn't be modifying it. you should be making money off it :) | 03:37 |
| SpamapS | patdk-lap: most of the time giving google a dedicated server meant FASTER load times. :) | 03:41 |
| patdk-lap | heh | 03:41 |
| patdk-lap | I dunno, I seem to always get hit by google, msn, and yahoo at the same time :( | 03:41 |
| SpamapS | patdk-lap: and since 90% of the pages would give back a 304 ... they wouldn't even consider the cold, unchanged reads. :) | 03:41 |
| patdk-lap | doesn't really bother me | 03:42 |
| lifeless | SpamapS: I wonder if I can impose on you to try a lucid rebuild | 03:42 |
| SpamapS | Though I do have to admit.. we had nearly 100GB of memcached available.. | 03:42 |
| lifeless | SpamapS: while I know how, I'm primarily focused on driving timeouts down | 03:43 |
| SpamapS | lifeless: sure, if it works I'll propose it as a lucid backport | 03:43 |
| lifeless | SpamapS: \o/ thanks | 03:43 |
| SpamapS | lifeless: as I'd expect.. latest haproxy builds fine on lucid.. can you give me the 2 line summary for why it is worth backporting to lucid so I can put it in the lucid-bacports request? | 03:54 |
| lifeless | SpamapS: to get option http-server-close | 03:55 |
| lifeless | http://haproxy.1wt.eu/download/1.4/doc/configuration.txt section 1.1 | 03:55 |
| patdk-lap | depressing, other than the usuals (navy.mil, freebsd.net, ntp.org), I have receive 1 email via ipv6, and it was spam | 03:57 |
| patdk-lap | in the last month | 03:57 |
| SpamapS | lifeless: oh.. I always sort of assumed haproxy worked that way ;) | 03:57 |
| SpamapS | patdk-lap: good to know the spammers will follow us into the future of the internets | 03:58 |
| lifeless | SpamapS: indeed, you can imagine our unpleasant surprise when we lowered the maxconn on edge 30 20 hours ago | 03:58 |
| lifeless | s/30 // | 03:58 |
| SpamapS | lifeless: bug #728204 has the backports request | 04:04 |
| uvirtbot | Launchpad bug 728204 in lucid-backports "Please backport haproxy 1.4.8-1 to lucid (currently 1.3.22-1)" [Undecided,New] https://launchpad.net/bugs/728204 | 04:04 |
| * SpamapS ponders whether or not to try and get another 30 minutes of sleep before his body notices the sun has started creeping over the horizon.. | 04:06 | |
| * patdk-lap makes spamaps work on the bug he found :) | 04:06 | |
| lifeless | SpamapS: I think you should sleep if you can | 04:06 |
| dravekx | zzzz.... updates | 04:10 |
| SpamapS | lifeless: I got 5 whole hours... who needs more than that? | 04:12 |
| * SpamapS will pay dearly for that at around 3:00pm today.. ;) | 04:13 | |
| lifeless | SpamapS: indeed | 04:13 |
| dravekx_ | how can I tell the version of a program installed via command line? | 05:37 |
| jmarsden | dravekx_: dpkg -l PACKAGENAME # If you don't know the PACKAGENAME, try dpkg -S `which PROGRAM` | 05:40 |
| dravekx_ | ok | 05:49 |
| dravekx | anyone know of a good free online server security check? | 06:56 |
| twb | Define "check" | 06:57 |
| dravekx | twb, uhm... | 06:59 |
| dravekx | check = "an online security scanner to see if there are any holes in your server security" ?? | 07:00 |
| twb | At layer 3, or 7? | 07:00 |
| dravekx | 7 | 07:01 |
| twb | No idea, sorry. | 07:01 |
| twb | For 3 I would run nmap or openvas from an endpoint outside your LAN | 07:02 |
| twb | For 7 I simply forbid users from running PHP. | 07:02 |
| dravekx | twp, that's a little ruff when msot of your users are wordpress and phpbb fanatics. | 07:02 |
| dravekx | most* | 07:02 |
| twb | Fuck em | 07:02 |
| dravekx | lol | 07:02 |
| twb | I would replace the former with something that generates static content (bloxsom?) and the latter with NNTP. | 07:04 |
| twb | Yeah, "Blosxom can operate as a CGI script or produce static HTML files." | 07:05 |
| twb | It's quite a lot harder to exploit bugs in sendfile(2) than in a PHP webapp | 07:05 |
| twb | Anyway, if all you care about is wordpress and phpbb, you could look at MITRE CVEs to learn about vulnerabilities. Combine that with whatever best current practices that upstream recommends, and you will be about as secure as you can get for those two | 07:09 |
| dravekx | twb, openvas is what i was looking for. works perfect. thanks. | 07:42 |
| twb | Well, I guess technically that's layer 7 | 07:43 |
| twb | But it'll test, say, apache2 rather than the apps running inside it | 07:44 |
| === ry_ is now known as ry | ||
| tap-out | hi, i ryn this coomand | 08:25 |
| tap-out | srry, quotacheck -avugm but it ask for quota force check ? | 08:26 |
| tap-out | any thingggggggggggggggggggggggggggggggggg | 08:30 |
| tap-out | i get this error wen i run quotacheck -avugm and it ask for quota force check | 08:31 |
| twb | tap-out: what is the full error? | 08:32 |
| tap-out | quotacheck -avugm , quotacheck: Quota for user is enabled on mountpoint / so might damage the file,please trun quota off or use -f to force checking | 08:33 |
| tap-out | twb, | 08:34 |
| twb | tap-out: it wants you to run "quotaoff /" | 08:34 |
| twb | tap-out: preferably after kicking all the users off the root filesystem | 08:34 |
| tap-out | twb, how i do that | 08:35 |
| twb | tap-out: quotacheck can only operate if the filesystem's quota isn't actively being updated | 08:35 |
| tap-out | u mean there is other user using it | 08:35 |
| tap-out | i am just using terminal and no other one in the server may some service | 08:36 |
| twb | tap-out: run "quotaoff /" | 08:36 |
| tap-out | how can i make quota off | 08:36 |
| twb | By fucking typing the command I just gave you | 08:37 |
| * twb rails at the language barrier | 08:37 | |
| tap-out | u fck command not work | 08:38 |
| twb | tap-out: did it give an error? | 08:39 |
| tap-out | quota -f , no file system specified | 08:39 |
| tap-out | twb, quota -f , no file system specified | 08:40 |
| twb | Sorry, you've exceeded my stupid threshold. Please wait for someone more tolerant to help you. | 08:40 |
| uvirtbot | New bug: #728299 in samba (main) "package samba 2:3.5.6~dfsg-5ubuntu2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728299 | 09:02 |
| _ruben | wow .. how can someone turn "quotaoff /" into "quota -f" and still able to operate a computer .. *amazing* | 09:08 |
| twb | _ruben: thrice, no less | 09:29 |
| phretor | I've noticed that Ubuntu redirects MySQL's error log to syslog. Is it possible to change this behavior? | 09:45 |
| twb | Probably | 09:46 |
| twb | "If you use mysqld_safe to start mysqld, mysqld_safe arranges for mysqld to write error messages to a log file or (as of MySQL 5.1.20) to syslog:" | 09:46 |
| twb | Looks like upstream changed the default: http://dev.mysql.com/doc/refman/5.1/en/error-log.html | 09:46 |
| twb | I imagine you want to pass --skip-syslog somewhere. | 09:47 |
| === twister004_ is now known as twister004 | ||
| uvirtbot | New bug: #728328 in nagios-plugins (main) "nagios plugins depend on samba" [Undecided,New] https://launchpad.net/bugs/728328 | 09:56 |
| === croessner is now known as croessner_ | ||
| === croessner_ is now known as croessner | ||
| === twister004_ is now known as twister004 | ||
| jibel | jamespage, Daviey can you help to test "UEC install (minimal topology)" amd64/i386 on Natty Alpha3 ? | 10:38 |
| Daviey | jibel, That is my plan :) | 10:38 |
| Daviey | jibel, hggdh and me were testing the candidate last night | 10:39 |
| jibel | Daviey, Awesome \o/ | 10:39 |
| === ogra_ is now known as ogra | ||
| uvirtbot | New bug: #728347 in squid (main) "package squid 2.7.STABLE7-1ubuntu12.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728347 | 10:46 |
| jibel | Anyone to test Install (default + RAID1) ? That's the last mandatory test for server on natty alpha3 | 10:48 |
| Daviey | jamespage, Are you snowed at the moment? | 10:58 |
| jamespage | working on the ec2 testing stuff at the moment | 10:58 |
| jamespage | Want me to pickup that test? | 10:58 |
| uvirtbot | New bug: #728362 in eucalyptus (main) "instance-id meta-service url 10.04.2" [Undecided,New] https://launchpad.net/bugs/728362 | 11:11 |
| Daviey | jamespage, if you could, that would be awesome... I'm kinda tied up on awstrial atm, causing pain | 11:52 |
| Daviey | smoser, When you are around - would you consider the upgrade delta warning for grub.conf, when upgrading grub a bug? | 11:58 |
| smoser | yes. | 12:00 |
| smoser | it forces a prompt ? | 12:00 |
| smoser | please open a bug. | 12:00 |
| Daviey | smoser, ack | 12:03 |
| jamespage | Daviey: won't get to it for an hour or so - is that OK? | 12:04 |
| Daviey | jamespage, Sounds good to me... jibel ? | 12:05 |
| * Daviey ^5's jamespage | 12:05 | |
| Daviey | jamespage, Actually, i could test it in VM. | 12:06 |
| jamespage | I will be | 12:06 |
| Daviey | jamespage, If i do amd64 now, can you do i386 in an hour or so? | 12:07 |
| smoser | i would like people's input on bug 728167 | 12:14 |
| uvirtbot | Launchpad bug 728167 in cloud-init "failure on apt-get update should be warning" [Medium,Triaged] https://launchpad.net/bugs/728167 | 12:14 |
| smoser | mainly, should failure of 'apt-get update' result in not running 'apt-get install <packages here>' where the user specified packages to run. | 12:14 |
| smoser | or should it just continue on, and warn that the update failed. | 12:14 |
| Daviey | smoser, I don't fully understand the issue... | 12:37 |
| Daviey | smoser, "the exception raised on failed 'apt-get upgrade' inside the cc_apt_update_" Is that a typo? s/upgrade/update/g' ? | 12:38 |
| smoser | yes | 12:39 |
| smoser | typo | 12:39 |
| Daviey | ahh | 12:39 |
| smoser | and now fixed | 12:40 |
| Daviey | So if it fails to update the cache, still try and install packages? | 12:40 |
| smoser | thank you for reading | 12:40 |
| Daviey | smoser, If that is so.. on the cloud images there is a pretty good chance the cache will be stale, compared to a traditional server install. | 12:42 |
| Daviey | marking the packages for installation might be a better solution? | 12:42 |
| m_tadeu_ | I'm trying to run a command in a remote server and, dispite I'm sure the file is there, it says "no such file or directory". What is happening in this case? | 12:42 |
| smoser | Daviey, the cache will be stale, but not invalid | 12:43 |
| smoser | oh wait.. yeah i fthere is a package in -updates you might be giong ot fail that too | 12:43 |
| Daviey | smoser, True, but what happens if the packages are not still in the pool? | 12:43 |
| Daviey | smoser, I wonder if the cloud images initial apt cache should be based on the release pocket only? | 12:44 |
| smoser | well, i was thinking that you'd get the one form the original release (not -updates) | 12:44 |
| smoser | but you're right | 12:44 |
| smoser | so what are you suggesting ? | 12:45 |
| Daviey | If the shipped cache is of the release pocket, install should always work | 12:45 |
| Daviey | If apt update works, great | 12:45 |
| Daviey | else they'll get the option next time they do update | 12:45 |
| Daviey | There is a risk they'll not get -security packages on first run though | 12:46 |
| Daviey | m_tadeu_, need more information to help really | 12:47 |
| Daviey | m_tadeu_, what command are you trying to run? | 12:48 |
| m_tadeu_ | Daviey: I'm accessing a remote server using ssh. I have a binary in /usr/local/fop2/fop2_server...when I "sudo /usr/local/fop2/fop2_server" it says file does not exist | 12:48 |
| m_tadeu_ | I even tryed to go to the dir and run it there....the file is there but bash says it doesn't exist | 12:49 |
| smoser | Daviey, shipped cache is of release pocket + point in time snapshot of -updates and -security | 12:49 |
| Daviey | m_tadeu_, ls -al /usr/local/fop2/fop2_server <--- that shows it? | 12:49 |
| m_tadeu_ | keep in mind that I'm even using tab for autocomplete | 12:50 |
| m_tadeu_ | Daviey: -rwxr-x--x 1 root root 6585775 2011-03-03 12:06 fop2_server | 12:50 |
| Daviey | smoser, That is my point... release pocket packages will be there for the life of the image.. -updates and -security is variable | 12:50 |
| smoser | right. | 12:51 |
| smoser | but what happens if | 12:51 |
| smoser | *package in -release | 12:51 |
| smoser | * package in -update [in cache] | 12:51 |
| Daviey | m_tadeu_, That is _odd_ | 12:51 |
| smoser | * newer package in -update [not in cache], so referenced deb is now gone | 12:51 |
| smoser | * apt-get install package | 12:52 |
| smoser | i think it fails because it sees that the version in its cache in -updates is newer than in -release | 12:52 |
| Daviey | m_tadeu_, Are you sure that is the file that can't be found, or is that trying to execute another binary? | 12:52 |
| smoser | and tries to get it | 12:52 |
| Daviey | smoser, hmm | 12:52 |
| dravekx | can someone explain to me how to create a cron job? I have one, but not sure how to make it run. | 12:52 |
| hallyn | ivoks: bug 726461 sure sounds familiar :) do you have anyone using ocfs on lucid who does not have instability? | 12:52 |
| Daviey | smoser, Hang on.. | 12:52 |
| uvirtbot | Launchpad bug 726461 in openssh "sshd on lucid causes kernel panic" [High,Incomplete] https://launchpad.net/bugs/726461 | 12:52 |
| Daviey | smoser, If you shipped the apt cache of just the release pocket, the packages in the pool should be there for the life of the image. | 12:53 |
| ivoks | hallyn: i have no problems | 12:53 |
| m_tadeu_ | Daviey: I have no idea, since it's a binary file, not a script. but I'm guessing if it would run another file it would be on the same dir....and now I'm on that dir | 12:53 |
| ivoks | hallyn: let me see :) | 12:53 |
| smoser | Daviey, right. | 12:53 |
| hallyn | ivoks: really? | 12:53 |
| smoser | but the apt-cache is just the apt-cache that the build was done on | 12:53 |
| hallyn | ivoks: ok, then maybe i'm just unfairly prejudiced | 12:53 |
| smoser | i *could* selectively clean it, yes. | 12:53 |
| Daviey | smoser, Yeah... that is what i was thinking | 12:54 |
| smoser | oh | 12:54 |
| smoser | shoot | 12:54 |
| smoser | heres why that doesn't work | 12:54 |
| Daviey | smoser, so you attempt to update, but if that fails you then know you can still install packages that will succeed | 12:54 |
| smoser | if you do that, and then run 'ubuntu-bug installed-package' where installed-package came from -updates | 12:54 |
| smoser | it will say "not an ubuntu package" | 12:54 |
| smoser | and you're confused | 12:54 |
| ivoks | hallyn: lol... this does look familliar, but i wouldn't blame ocfs2 yet | 12:55 |
| Daviey | m_tadeu_, Sorry, i have no experience with that binary. The best i can direct you to is googleing for that binary name and the error message | 12:55 |
| Daviey | m_tadeu_, hopefully that will help | 12:55 |
| Daviey | smoser, hmm | 12:55 |
| m_tadeu_ | Daviey: tryed that before comming here :) | 12:55 |
| Daviey | smoser, hmm | 12:56 |
| ivoks | hallyn: i've had problems with iscsi, though | 12:56 |
| Daviey | smoser, I guess this matters more for post release images, rather than natty final | 12:56 |
| smoser | Daviey, so i actually had to do some work to make this all work the way it does... we rename the cache files so they match the mirror that we select on boot. | 12:56 |
| Daviey | smoser, BUT, this is just to cater for failure when initial update fails | 12:57 |
| smoser | yes. | 12:57 |
| smoser | so maybe if update fails, just tag for isntallation and warn | 12:57 |
| smoser | but what sucks about that is that in most cases, it *would work*. (most packages don't have updates in -updates) | 12:58 |
| Daviey | smoser, yeah.. i don't know how to mark for installation in apt-get, but aptitude is easy to do that. | 12:58 |
| ivoks | hallyn: otoh, do_filp_open is a function in fs/ | 12:58 |
| smoser | surely there is some way. i will just dig a bit. | 12:58 |
| smoser | thanks. | 12:58 |
| Daviey | smoser, Keep us posted on what you do. | 12:59 |
| hallyn | ivoks: in the past that sort of oops signature wouldve convinced me aufs was to blame, but that seems not involved here | 13:00 |
| smoser | oh yeah, i'm sure you're on the edge of your seat :) | 13:00 |
| smoser | the real key is for the update to not fail | 13:00 |
| zul | hmmmm? | 13:00 |
| smoser | (and it shouldn't we've got local mirrors) | 13:00 |
| smoser | but it could still fail in the case of user added repositories | 13:01 |
| Daviey | smoser, Good point | 13:03 |
| smoser | i'm thinking try and fail | 13:04 |
| smoser | and warn | 13:04 |
| zul | whats going on about? | 13:04 |
| Daviey | smoser, easiest! | 13:05 |
| Daviey | zul, see scrollback | 13:05 |
| jibel | Daviey, jamespage sounds good to me too, thanks. | 13:07 |
| ivoks | hallyn: yeah... but i don't believe it's ocfs; i refuse to believe :D | 13:09 |
| jamespage | Daviey: ack | 13:23 |
| hallyn | ivoks: well, the thing is, the idiom where putfile calls get_empty_filp() doesn't really seem like it fits anywhere but aufs | 13:29 |
| === taneli_ is now known as xampart | ||
| ivoks | hallyn: maybe he uses aufs :) | 13:33 |
| hallyn | oh no i was misreading | 13:36 |
| uvirtbot | New bug: #728442 in squid (main) "package squid 2.7.STABLE7-1ubuntu12.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728442 | 13:36 |
| ivoks | hallyn: well, if i understand correctly, he's not using ocfs2 on the box that crashes | 13:37 |
| ivoks | hallyn: he's just doing lvm snapshot | 13:37 |
| ivoks | oh, it does mount it, but after the snapshot is taken | 13:38 |
| zul | uvirtbot: yes we know | 13:40 |
| uvirtbot | zul: Error: "yes" is not a valid command. | 13:40 |
| uvirtbot | New bug: #728444 in squid (main) "package squid 2.7.STABLE7-1ubuntu12.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728444 | 13:41 |
| RoAkSoAx | morning all | 13:43 |
| RoAkSoAx | hggdh: you ready? | 13:44 |
| KurtKraut | How do I remove a package intentionally breaking its dependencies? I want to remove the most to date update version of libapache2-mod-wsgi without removing the packages that depends on in and them manually install thru dpkg a .deb file of an old version of libapache2-mod-wsgi. How do I remove it? | 13:45 |
| hggdh | RoAkSoAx, I am the readiest I can ever be ;-) | 13:45 |
| hggdh | and good morning | 13:45 |
| RoAkSoAx | hggdh: morning ;) | 13:46 |
| RoAkSoAx | hggdh: alright fire everything up | 13:46 |
| ivoks | KurtKraut: dpkr -r --ignore-depends=<package>,... | 13:46 |
| ivoks | KurtKraut: dpkg, of course :) | 13:46 |
| RoAkSoAx | hggdh: i'll prepare a config for you to use. btw. Those servers support suspending right? | 13:47 |
| ivoks | KurtKraut: or... dpkg -r --force-depends libapache2-mod-wsgi | 13:47 |
| hggdh | RoAkSoAx, I believe they do, never had to try it before | 13:47 |
| hggdh | RoAkSoAx, and, right now, they are all under powernap control and grasp | 13:48 |
| KurtKraut | ivoks, it worked, thanks! | 13:48 |
| RoAkSoAx | hggdh: what are we gonna do is change the config to how it was originally and see what happens | 13:49 |
| Disconnect | anyone else running into an issue where you can't enable pxe at all on kvm vms? if pxelinux loads (even with localboot 0, localboot -1, localboot 0x80) it says "Booting from local hard drive" and then cycles immediately back to dhcp/netbooting.. I'm suspecting its cuz they use virtio but if I cancel netbooting before that ('n' to netboot, 'q' to quit) they boot fine.. | 13:49 |
| RoAkSoAx | hggdh: please change the config on the servers running powernap to http://pastebin.ubuntu.com/574975/ | 13:49 |
| RoAkSoAx | and sudo service powernap stop && sudo service powernap start again to make sure that the changes made effect | 13:50 |
| RoAkSoAx | hggdh: and then tail -f /var/log/powernap.log | 13:50 |
| RoAkSoAx | and let's let eucalyptus figure out what to do | 13:50 |
| hggdh | ok | 13:52 |
| hggdh | RoAkSoAx, interesting. The machines are up, but unresponsive | 13:54 |
| RoAkSoAx | hggdh: what should happen though is that powernap will always detect /sbin/init and will never do any action by itself. So, since eucalyptus decides to poweroff and stuff, eucalyptus will just send a command to the NC's and they will suspend of poweroff | 13:55 |
| RoAkSoAx | hggdh: any logs/ both eucalyptus/powernap? | 13:55 |
| hggdh | RoAkSoAx, not right now, having to reboot the CLC | 13:56 |
| hggdh | RoAkSoAx, the CLC is unresponsive | 13:56 |
| RoAkSoAx | oki | 13:56 |
| hggdh | RoAkSoAx, I am assuming that all other config options are good as they are now | 13:59 |
| hggdh | and I have restarted powernap on the NC, waiting for the CLC to finish reboot | 13:59 |
| RoAkSoAx | hggdh: ok ;) | 14:03 |
| kriswalo | Hi, my ubuntu server "crashed"/froze/"disapered from the network" a couple of days ago so I had to physically reboot it to be able to reconnect to it. Now I got my connection back up, but I'm not able to run a python script I have that's supposed to run and manage transmission for me. I get 401: Unauthorized user. Can someone help me? | 14:04 |
| hggdh | RoAkSoAx, OK. I have now to restart the NC, since whatever happened has already happened -- the CLC cannot see any regions available on the NC | 14:04 |
| RoAkSoAx | hggdh: ok so I'm assuming that PowerNap itself has not suspeneded the NC's since the ProcessMonitor tracks the init daemon which will always be present in the process table. | 14:05 |
| RoAkSoAx | hggdh: can you show me powernap logs from any NC aswell as eucalyptus logs where it shows what it is doing with powernap (if it is sending wakeup signals and stuff) | 14:05 |
| hggdh | RoAkSoAx, I think it had done so _before_ I restarted it, and the euca tables got confused | 14:05 |
| hggdh | RoAkSoAx, will do, as soon as system is back up | 14:06 |
| RoAkSoAx | ok | 14:06 |
| patdk-wk | still working on that? | 14:06 |
| RoAkSoAx | patdk-wk: wer are just about to start | 14:07 |
| hggdh | RoAkSoAx, well, guess what... | 14:12 |
| hggdh | RoAkSoAx, now the CLC does not contact the NC | 14:12 |
| hggdh | not even to start | 14:12 |
| ivoks | hallyn: http://oss.oracle.com/pipermail/ocfs2-users/2010-September/004686.html | 14:12 |
| RoAkSoAx | hggdh: can you ssh into the NC? | 14:13 |
| hggdh | RoAkSoAx, yes. But Euca (on the NC) is reporting apache2: symbol lookup error: /etc/eucalyptus/axis2/services/EucalyptusNC/libEucalyptusNC.so: undefined symbol: rampart_print_security_processed_results_set | 14:13 |
| RoAkSoAx | hggdh: ahh!1 that's the same error I was experiencing. But, wasn't that fixed already? | 14:14 |
| hggdh | so we cannot really test -- we need the NC to register sanely at least once :-( | 14:14 |
| RoAkSoAx | Daviey: ^^ | 14:14 |
| hggdh | RoAkSoAx, seems to be new, actually, I did not see it on A2 | 14:14 |
| hallyn | ivoks: (curse needless line wrapping) that one looks like memory got walked over | 14:14 |
| hggdh | (new == new since A2) | 14:14 |
| hggdh | GOOD MORNING Daviey | 14:15 |
| ivoks | hallyn: yeah, it's not the same | 14:15 |
| RoAkSoAx | hggdh: can you ssh into the NC?724470 | 14:15 |
| RoAkSoAx | argh | 14:15 |
| RoAkSoAx | bug | 14:15 |
| RoAkSoAx | bug #724470 | 14:15 |
| uvirtbot | RoAkSoAx: Error: Could not parse data returned by Launchpad: The read operation timed out | 14:16 |
| hallyn | but how come it *always* has that do_filp_open -> get_empty_filp -> apparmor_alloc_file_security trace (-> kzalloc) trace? | 14:16 |
| hggdh | RoAkSoAx, yes, I can ssh in | 14:16 |
| RoAkSoAx | hggdh: the bit ius wrong: check https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/724470 | 14:16 |
| uvirtbot | Launchpad bug 724470 in eucalyptus "apache2: symbol lookup error: /etc/eucalyptus/axis2/services/EucalyptusNC/libEucalyptusNC.so: undefined symbol: rampart_print_security_processed_results_set" [High,Fix released] | 14:16 |
| hallyn | ivoks: and have you ever seen this with 2.6.35? of course people will use this in LTS which is .32, but I do wonder if it's been long fixed | 14:16 |
| ivoks | hallyn: i haven't see it with .32 either | 14:17 |
| hggdh | RoAkSoAx, yes, I am running rev1255, will update | 14:18 |
| hallyn | ivoks: you mean you, personally, have not see it on .32, right? | 14:18 |
| Daviey | morning hggdh | 14:18 |
| Daviey | hggdh, that should be fixed!! | 14:18 |
| hallyn | Daviey: do you have any idea (a) what's up with and (b) if anyone's looking at bugs 727342 and 727586 (and the several other similar ones)? | 14:18 |
| uvirtbot | Launchpad bug 727342 in open-vm-tools "package open-vm-dkms 2010.06.16-268169-3ubuntu1 failed to install/upgrade: open-vm-tools kernel module failed to build" [Undecided,New] https://launchpad.net/bugs/727342 | 14:18 |
| uvirtbot | Launchpad bug 727586 in open-vm-tools "package open-vm-dkms 2010.06.16-268169-3ubuntu1 failed to install/upgrade: open-vm-tools kernel module failed to build" [Undecided,New] https://launchpad.net/bugs/727586 | 14:19 |
| hallyn | Daviey: or rather, does anyone look after open-vm-tools? | 14:19 |
| Daviey | hallyn, I don't think anyone is looking at those atm | 14:19 |
| hallyn | i think for starters I'll mark them all dups and confirmed... | 14:19 |
| hggdh | Daviey, it probably is, but not on a3 (as yesterday' s image had it) | 14:19 |
| ivoks | hallyn: right | 14:19 |
| hggdh | Daviey, I am upgrading to rev1256 | 14:19 |
| Daviey | hggdh, Looks like it's been community maintained | 14:20 |
| hallyn | ivoks: I supposed it's not unlike those weird bridging bugs - too dependent on environment to reproduce :( | 14:21 |
| Daviey | hallyn, I just messaged the last uploader, and the person that sponsored it to see if they want to touch it | 14:22 |
| ivoks | hallyn: we could try reproducing it, but i don't have hardware atm :) | 14:22 |
| Daviey | hallyn, sadly, the person that last changed it is offline | 14:22 |
| jibel | jamespage, Daviey , hggdh ,can one of you pickup the test "UEC install (minimal topology)" ? | 14:22 |
| hallyn | Daviey: oh, cool, thanks. Though my guess is that it's not a regression from a bad upload, but just out of date kernel module code | 14:22 |
| Daviey | hallyn, yeah | 14:23 |
| Daviey | jibel, I think hggdh is already hammering that | 14:23 |
| Daviey | hallyn, I haven't compared but it could be the same issue as bug #498677 | 14:24 |
| uvirtbot | Launchpad bug 498677 in open-vm-tools "open-vm-tools kernel modules don't build on lucid kernels" [Medium,Fix released] https://launchpad.net/bugs/498677 | 14:24 |
| Daviey | hallyn, which could mean, as you say - it needs refreshing against the latest kernel | 14:24 |
| hggdh | I had forgotten to update the status of the test, jibel | 14:24 |
| hallyn | Daviey: pls ping me if you hear back from the uploader | 14:25 |
| Daviey | hallyn, Which means an upstream fix (they fixed the last one). | 14:25 |
| Daviey | hallyn, Have you compared upstream commits looking for a fix? | 14:26 |
| hallyn | Daviey: no, i'm wasting my f'ing time marking 100 f'ing dupes | 14:26 |
| Daviey | hallyn, :(... I think we need an LP API tool for that. | 14:26 |
| hallyn | Daviey: well, we need a maintainer for that package so that the bugs don't stack up until my 'why do i see a new bug like this every day in my inbox' filter goes off :) | 14:27 |
| Daviey | heh | 14:27 |
| hggdh | Daviey, what is the issue on the upstart failing on euca? | 14:27 |
| hggdh | just got ity | 14:27 |
| Daviey | hggdh, It's eucalyptus-network upstart job blocking | 14:28 |
| hggdh | Daviey, any bypass? | 14:28 |
| Daviey | hggdh, not exactly sure yet... it's either a change in upstart or eucalyptus-networking job failing. | 14:29 |
| Daviey | perhaps due to dhcpd | 14:29 |
| Daviey | hggdh, the bypass is simple... | 14:29 |
| Daviey | hggdh, What server is this on? | 14:29 |
| hggdh | Daviey, sapodilla and soncoya | 14:30 |
| Daviey | hallyn, I'm feeling the dupes on bug #717397 | 14:30 |
| uvirtbot | Daviey: Error: Could not parse data returned by Launchpad: The read operation timed out | 14:30 |
| hallyn | Daviey: well while there are bugs for lucid, mav, and natty, i think each will need a different fix | 14:32 |
| hallyn | not sure though. almost ready to d/l the source :) | 14:32 |
| hggdh | Daviey, bypassed, quite easy. It is a change in upstart | 14:32 |
| * SpamapS perks up | 14:33 | |
| SpamapS | upstart? | 14:33 |
| hallyn | oh no, we woke up SpamapS | 14:33 |
| Daviey | hggdh, http://pb.daviey.com/HiVP/raw/ | 14:33 |
| hallyn | SpamapS: no no, we were just talking about sysvinit and how it rocks | 14:33 |
| SpamapS | hallyn: +1 ;) | 14:33 |
| jkg | all this new-fangled upstart stuff, much like modern hardware/device management, and things like apparmor, make me feel like a bit of a luddite :-/ | 14:34 |
| hggdh | RoAkSoAx, now we are ready | 14:37 |
| RoAkSoAx | hggdh: alrighty | 14:37 |
| Daviey | hggdh, Pain that euca didn't make the Alpha cd. | 14:40 |
| hggdh | RoAkSoAx, here's what I will do: I will start submitting instances; they all should fail (because of the DHCP issue) | 14:40 |
| Daviey | Yeah | 14:40 |
| RoAkSoAx | ok | 14:40 |
| hggdh | Daviey, yeah :-) | 14:40 |
| hggdh | RoAkSoAx, so we should have the NC doing the same thing as before | 14:40 |
| Daviey | oh goody | 14:41 |
| Daviey | hallyn, the person that sponsored it wasn't involved in fixing it | 14:43 |
| RoAkSoAx | hggdh: so this time the NC should go to sleep and then when trying to fire up an instance it should come back up | 14:43 |
| hggdh | yes. Let's hope it does get to the 'when trying to fire up an instance' -- which I am not really sure right now | 14:45 |
| iggi | Question, is there anything like oVirt for an Ubuntu machine? | 14:47 |
| Daviey | soren, Didn't you play with ovirt ages ago? | 14:48 |
| iggi | Daviey: I did see something on the list, but it was circa 2008 | 14:49 |
| iggi | The main problem is that the only competitor I can find went closed source (enomolism?) | 14:50 |
| soren | Daviey: No, I decided it wasn't interesting enough to sustain The Pain of A Million Ruby Gems[tm]. | 14:50 |
| soren | Daviey: Or thereabouts. I didn't actually count them. | 14:50 |
| iggi | soren: heh, fair enough. I will have to live with virt-manager+x11 forwarding unless you know a better way | 14:51 |
| hallyn | Daviey: (just curious) meaning he got back to you and said that, or that you thought I was casting random blame left and right? :) | 14:51 |
| soren | iggi: Why do X11 forwarding? | 14:52 |
| hallyn | Daviey: I'm having some trouble with the build-depends being insufficient, tbh. | 14:52 |
| hallyn | hm, or, no, something went wrong on thsi system. feh. | 14:52 |
| iggi | soren: virt-manager and connecting to nodes turned out to be a pain last I tried. I could only get it working via SSH and that required a password for every action. I supposed PSK, but never had that working right with virt-manager. | 14:53 |
| iggi | suppose* | 14:53 |
| soren | iggi: ssh forwarding is by far the easiest. It automatically tunnels vnc connections, too. | 14:54 |
| soren | It's quite simple to set up, really. ssh-keygen, ssh-copy-id, done. | 14:54 |
| RoAkSoAx | hggdh: anything? | 14:56 |
| hggdh | RoAkSoAx, differently from the previous runs, it now *never* reports 0 resources | 14:57 |
| hggdh | RoAkSoAx, so I canceled the test, and am now giving it a time for powernap to kick in | 14:58 |
| RoAkSoAx | hggdh: ok, so that's good I pressume, since it will wakeup the NC whenever an instance is requested to run | 14:58 |
| hggdh | RoAkSoAx, actually, I would have to restart the NC -- the instances are now stuck in shutting down | 14:59 |
| hggdh | done. Now we wait for powernap to kick in (again) | 15:01 |
| * kirkland high fives hggdh and RoAkSoAx | 15:02 | |
| kirkland | powernaps to the people! | 15:02 |
| RoAkSoAx | kirkland: hurray!! | 15:02 |
| Daviey | hallyn, nah - seems to have been a drive by sponsor. | 15:02 |
| iggi | Bah, cPanel does not support Ubuntu? :( | 15:02 |
| hggdh | heh | 15:03 |
| RoAkSoAx | hggdh: anything yet? | 15:14 |
| b0gatyr | hi everyone | 15:15 |
| hggdh | RoAkSoAx, it seems powernap never kicked in | 15:17 |
| RoAkSoAx | hggdh: powernap by itself doesn't do anything afaik | 15:18 |
| RoAkSoAx | kirkland: ping | 15:18 |
| RoAkSoAx | hggdh: did it use to work out of the box before? | 15:19 |
| hggdh | RoAkSoAx, this is similar to what I had before -- I never saw anything different | 15:19 |
| RoAkSoAx | hggdh: for what I know the CC (or CLC) automatically decide what machine to powernap/wakeup | 15:20 |
| RoAkSoAx | so powernap always runs | 15:20 |
| RoAkSoAx | tracking sbin/init which will always be running | 15:20 |
| RoAkSoAx | so powernap by itself will never perform an action | 15:20 |
| RoAkSoAx | so when the cloud decices which NC to powernap, it sends a powernap-now command to the NC, which execs through the powernapd daemon, and performs the action (i.e. suspend) | 15:21 |
| RoAkSoAx | and to wakeup it just sends a WoL | 15:21 |
| RoAkSoAx | can I see your logs though? | 15:21 |
| hggdh | RoAkSoAx, which logs you want? | 15:30 |
| RoAkSoAx | hggdh: /var/log/powernap.log | 15:30 |
| RoAkSoAx | hggdh: and I believe it is the NC log which shows when eucalyptus tris to wkaeup/powernap the NC's | 15:30 |
| hggdh | RoAkSoAx, powernap.(err|log) are empty | 15:32 |
| hggdh | RoAkSoAx, http://people.canonical.com/~cerdea/nc.log | 15:34 |
| RoAkSoAx | hggdh: that's weird, did since DEBUG=3 in the powernap config | 15:34 |
| hggdh | RoAkSoAx, duh, I had debug=1 | 15:35 |
| hggdh | restarted powenap now | 15:36 |
| RoAkSoAx | hggdh: did you use the config settings I gave you? | 15:36 |
| hggdh | RoAkSoAx, see above. All there, but debug was set to 1 by mistake | 15:37 |
| RoAkSoAx | hggdh: ok the log by no means shows that any NC is idled | 15:38 |
| RoAkSoAx | like it should at https://help.ubuntu.com/community/UEC/PowerManagement | 15:38 |
| bobg | I am trying to get sudo-ldap installed in a preseed file. It fails because the prerm script of sudo pkg refuses because there is no root password (it does not know that sudo-ldap will replace it) I have tried preseeding a root password, I have tried setting "export SUDO_FORCE_REMOVE=yes in a preseed early_command. How can I get sudo-ldap installed instead of sudo in a preseed installation? | 15:42 |
| _wo_ | hello | 15:42 |
| _wo_ | i dont hang out here | 15:42 |
| _wo_ | but im looking for info about running oprofile on amazon's ec2 platform | 15:43 |
| _wo_ | has anyone been able to do this? | 15:43 |
| bobg | _wo_, its ok, noone wants to admit hanging out here. The first step is to admit you hae a problem | 15:44 |
| _wo_ | hehe | 15:44 |
| _wo_ | i always feel like a jerk for bursting into a chan and demanding answers to my questions | 15:45 |
| hggdh | RoAkSoAx, even after resetting DEBUG to 3, and 'sudo restart powenap', the logs are still empty | 15:47 |
| hggdh | RoAkSoAx, OK. When nothing else works, power it off, count to 10, power it on again. | 15:49 |
| RoAkSoAx | hggdh: yeah you need to sudo service powernap stop / sudo service powernap start for the settings to make effect | 15:51 |
| RoAkSoAx | I think that's a bug in upstart, isn't it SpamapS ? | 15:51 |
| hggdh | RoAkSoAx, restarted | 15:52 |
| Disconnect | anyone know why d-i isn't accepting the hostname? (dns resolution doesn't work, hoping thats not it..) its set in the preseed, but it still prompts. if i set it in the pxe config it works, but I need to preseed it.. (worst case, I can write it to /etc/hostname before rebooting and let puppet take care of fixing /etc/hosts, but thats a hack and a half..) | 15:56 |
| RoAkSoAx | hggdh: man I now understand what you go through when testing Euca!! hehe | 15:57 |
| bobg | Disconnect, what's the exact d-i line you are using? | 15:57 |
| hggdh | RoAkSoAx, heh. Welcome to hell :-) | 15:57 |
| thesheff17 | Disconnect: d-i netcfg/get_hostname string hostname do you have this? | 15:57 |
| Disconnect | d-i netcfg/get_hostname string test-host & d-i netcfg/get_domain string domain.com | 15:58 |
| Disconnect | its popping up with 'ubuntu' as the default. | 15:58 |
| Disconnect | (fyi if anyone else is having issues with pxelinux not accepting localboot on kvm vms, chain.c32 works fine.) | 16:00 |
| RoAkSoAx | hggdh: brb gonna grab some chips and something to drink for today's long journey of testing euca+powernap :P | 16:00 |
| bobg | Disconnect, looks fine to me -- I don't set the hostname in preseed | 16:00 |
| Disconnect | the domain is set correctly in /etc/hosts (with no prompt) it just doesn't like the hostname :/ | 16:01 |
| Disconnect | once this build finishes i'll see what is in debconf. maybe there is something else related.. | 16:01 |
| bobg | Disconnect, do you install sudo-ldap in your preseed by any chance? (that's my problem -- I can't get that package to replace sudo in the preseed install) | 16:02 |
| Disconnect | yah i do in fact :) have to hack it with a post-inst script though. hang on i'll pastebin something for you | 16:03 |
| bobg | Disconnect, cool, you da' man (although I guess with the nickname Disconnect, I can't be sure :) | 16:04 |
| Disconnect | http://pastebin.com/vs1vXwjL | 16:05 |
| Disconnect | -might- be able to just "in-target env .."etc but I have a bunch of other post-run stuff | 16:05 |
| bobg | thanks a bunch -- I am off to try it | 16:08 |
| Disconnect | cool enjoy :) | 16:08 |
| bobg | Disconnect, I just started an install to test. I also set the hostname in the preseed, but it still prompted me too | 16:15 |
| Disconnect | awesome :/ at least i'm not alone :) | 16:15 |
| georgie_b | can any one tell me a real time web traffic monitor software for xampp | 16:17 |
| georgie_b | maybe also the cpu and memory usage load,performance monitors to0 | 16:18 |
| georgie_b | hello | 16:18 |
| georgie_b | ne body there? | 16:18 |
| PrickelPit | apachetop | 16:20 |
| georgie_b | @PrickelPit : is apachetop a free software? | 16:23 |
| Disconnect | damn. debconf-get-selections --installer | grep test- just shows the same hostname. time to dig through the file by hand.. | 16:26 |
| RoyK | georgie_b: bsd license afaics | 16:27 |
| RoAkSoAx | hggdh: im setting euca im my local machine | 16:31 |
| hggdh | RoAkSoAx, I now have entries in powernap.log, but nothing seems to have happened. You want the log? | 16:32 |
| RoAkSoAx | hggdh: sure. what intrigues me is why euca is not reporting anything | 16:33 |
| hggdh | RoAkSoAx, indeed. No euca logs have entries matching 'wake' or 'power' | 16:34 |
| hggdh | RoAkSoAx, http://people.canonical.com/~cerdea/powernap.log | 16:35 |
| RoAkSoAx | hggdh: yeah as expected powernap is running and just waiting for a command from euca | 16:36 |
| RoAkSoAx | which never comes | 16:36 |
| hggdh | RoAkSoAx, we can try now with the old settings, and see what happens | 16:43 |
| RoAkSoAx | ok | 16:43 |
| === mdeslaur_ is now known as mdeslaur | ||
| Disconnect | bobg: fyi if you add "d-i netcfg/confirm_static boolean true | 17:00 |
| Disconnect | then hostname preseeding works | 17:00 |
| bobg | Disconnect, thanks for the tip | 17:01 |
| bobg | Disconnect, btw, my install is getting close to the point where it fails to install the sudo-ldap pkg -- fingers crossed | 17:02 |
| Disconnect | good luck :) | 17:03 |
| === alex_jon1 is now known as alex_joni | ||
| Disconnect | oops maybe not. it worked once, then prompted. sigh. i'll keep arguing with it. | 17:07 |
| hggdh | RoAkSoAx, with ACTION_METHOD=0 the system went into powersave (so it seems). But euca did not trigger anything visibly | 17:20 |
| RoAkSoAx | hggdh: yeah so for euca the NC is still running | 17:21 |
| hggdh | indeed | 17:21 |
| bobg | Disconnect, my sudo-ldap pressed installed -- thanks | 17:22 |
| RoAkSoAx | hggdh: change ACTION_METHOD=1 and let the machine suspend and we'll see what euca does | 17:22 |
| hggdh | RoAkSoAx, ack | 17:22 |
| RoAkSoAx | hggdh: better yet,m change it to 4 | 17:22 |
| RoAkSoAx | in case it does not support suspend it will poweroff | 17:22 |
| hggdh | k | 17:23 |
| dmesg | all will be a | 17:24 |
| RoAkSoAx | hggdh: btw.. is the scehduler policy Round Robin or PowerSave? | 17:27 |
| hggdh | RoAkSoAx, roundrobin | 17:28 |
| RoAkSoAx | hggdh: ok so then if we are in roundrobin then powernap should not do any action on the NC's because that should be managed by euca when on SCHEDPOLICY=POWERSAVE | 17:30 |
| === jussi01_ is now known as jussi | ||
| Disconnect | bobg: cool no problem. (that is the biggest most non-debian/ubuntu mess i've ever seen. we can't install it with puppet because of the environment var. can't preseed it, can't install it from apt..) | 17:30 |
| === mrmist_ is now known as mrmist | ||
| hggdh | duh | 17:30 |
| Disconnect | rather, from aptitude (or apt-dater or ..) | 17:31 |
| RoAkSoAx | hggdh: so that means if running ROUNDROBIN then we should either disable powernap or configure it to monitor the Process "init" | 17:31 |
| Disconnect | can't sudo install it unless you let them use 'env'.. | 17:31 |
| RoAkSoAx | hggdh: so that it will never send the machine to "sleep" | 17:32 |
| hggdh | RoAkSoAx, let set it back to monitor init first | 17:32 |
| hggdh | RoAkSoAx, with action-method=4 | 17:33 |
| RoAkSoAx | hggdh: yes having that would be how powernap default before natty | 17:33 |
| RoAkSoAx | hggdh: | 17:34 |
| RoAkSoAx | [ProcessMonitor] | 17:34 |
| RoAkSoAx | #init = "^/sbin/init" | 17:34 |
| RoAkSoAx | just uncomment that line | 17:34 |
| RoAkSoAx | what will happen is that powernap will never go to sleep because it will always detect the init process | 17:34 |
| hggdh | RoAkSoAx, powernap restarted as above | 17:35 |
| === kriswalo is now known as eitri | ||
| RoAkSoAx | hggdh: ok so this way everything should be as original | 17:40 |
| RoAkSoAx | i mean the beahviour as un previos releases should not change | 17:40 |
| RoAkSoAx | hggdh: I'm gonna write a before/after doc explaining how it was working and how is working now that will help us understand the issues | 17:40 |
| === paul__ is now known as orudie | ||
| === Visig0th is now known as Visigoth | ||
| RoAkSoAx | hggdh: let me ask you on question though. If we are in ROUNDROBIN and you manually turn off an NC, will the CLC/CC try to wake it up or will just assume that the node is off/failed or not available | 17:44 |
| bobg | hey Disconnect, are you providing your preseed over tftp? Isn't it too late to provide the network configuration then? Maybe its ignoring all of the netcfg preseeds because it already configured the network to access tftp | 17:46 |
| Disconnect | it'll get it over http. in pxelinux append "url=http://(our foreman server):3000/unattended/preseed" | 17:47 |
| hggdh | RoAkSoAx, on round-robin it would assume the NC off | 17:47 |
| Disconnect | it says dhcp hostname will override the preseed if it is provided, but dhcp isn't providing a hostname.. (as evidenced by the fact that it takes domain from preseed and then prompts for the "missing" hostname..) | 17:48 |
| === hikeonpast__ is now known as hikeonpast | ||
| bobg | Disconnect, what I am suggesting is that maybe the installer inspects all of the netcfg/ debconf vars (even the hostname) when its first initializing the network. Then it loads your preseed from the network but it ignores the dbconf values that it has already queried | 17:51 |
| Disconnect | what i suspect its doing is: read netcfg from commandline, feed hostname/etc into dhcp client. get response back (including same hostname if present), use hostname, get preseed. if that hostname isn't present on the commandline, it'll come back blank and - for whatever reason - that overrides the debconf entry anyway.. | 17:55 |
| jfluhmann | running into a situation similar to bug 29858 (https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/29858) | 18:01 |
| jfluhmann | I have my root mounted on lvm that includes local drives and a Dell MD3000 storage array | 18:01 |
| uvirtbot | jfluhmann: Error: Could not parse data returned by Launchpad: The read operation timed out | 18:01 |
| uvirtbot | jfluhmann: Error: Could not parse data returned by Launchpad: The read operation timed out | 18:01 |
| jfluhmann | worked fine when using just the local drives | 18:01 |
| jfluhmann | system dropped to the busybox shell upon reboot after adding the MD3000 storage (which is similar to bug #29858) | 18:01 |
| uvirtbot | Launchpad bug 29858 in initramfs-tools "root on lvm fails to boot." [High,Confirmed] https://launchpad.net/bugs/29858 | 18:01 |
| jfluhmann | I'm able to exit the shell and go to my normal login and things seem to work | 18:01 |
| jfluhmann | any direction on where I can start looking? Seems like there's a step in the boot process that needs to be moved. | 18:01 |
| === JayFo is now known as JFo | ||
| === hikeonpast__ is now known as hikeonpast | ||
| Bilge | How can it be that if I set the owner to root and the chmod to '0' for a file | 18:26 |
| Bilge | i.e. nobody can read/write/execute it | 18:26 |
| Bilge | That a program can set the owner to itself and give itself read and write permissions | 18:26 |
| geekbri | any start-stop-daemon experts in here? No process in pidfile '/var/run/stage2_loader/reporting_stage2_loader.pid' found running; none killed. i'm getting that error message but when i cat the pidfile it does indeed contain the proper pid for the daemon running | 18:27 |
| === paul_ is now known as orudie | ||
| patrickw | Is anyone running OpenAM (server + agent) on Lucid? I can't seem to get it working... | 18:29 |
| patrickw | Err, and by OpenAM I meant the fork of Sun's OpenSSO, not the answering machine or whatever else is called OpenAM. | 18:30 |
| === cmagina is now known as cmagina-lunch | ||
| === nomadgis_ is now known as nomadgis | ||
| hggdh | RoAkSoAx, I do not see the symptoms I had on A2 | 18:59 |
| RoAkSoAx | hggdh: alrighty then so we can assume everything is working as expected. now the issue might appear when we use SCHEDPOLICY=POWERSAVE | 19:01 |
| RoAkSoAx | hggdh: explanation here: http://pastebin.ubuntu.com/575114/ | 19:01 |
| hggdh | RoAkSoAx, agreed | 19:04 |
| RoAkSoAx | hggdh: will email kirkland about it to decide default behavior and so that it remains compatible out of the box for eucalyptus | 19:05 |
| jiboumans | smoser, around? | 19:05 |
| jiboumans | i'm seeing this in my logs on a few machines; does this ring a bell to you? http://pastebin.com/BKHwHqDf | 19:06 |
| patdk-wk | jiboumans, google it | 19:06 |
| patdk-wk | that has been around forever | 19:06 |
| jiboumans | patdk-wk: that's exactly why i ask | 19:07 |
| jiboumans | there's tons of issues reported, and fixes as well | 19:07 |
| jiboumans | i'm curious as to which may or may not have made it to 10.04.2 | 19:07 |
| Daviey | jiboumans, Wow, that is a throwback from the past.. I thought that was dealt with now, what release is that? | 19:07 |
| jiboumans | 10.04.2, latest amis | 19:07 |
| Daviey | crikey | 19:07 |
| smoser | jiboumans, here. | 19:07 |
| jiboumans | the pids its reporting are no longer running on the machine, and i have been experiencing xvfb crashes on those, hence i'm wondering if its' related | 19:08 |
| smoser | jiboumans, 'import' like image magick? | 19:08 |
| jiboumans | ah hmm | 19:09 |
| jiboumans | smoser: yes, there's a cronjob that takes screenshots. i belive it uses image magick | 19:09 |
| jiboumans | smoser: reckon this issue can be taking down xvfb? | 19:09 |
| smoser | i suspect that binary is not built properly built. i havenot looked much at it | 19:10 |
| smoser | there is comment to that effect in https://bugs.launchpad.net/ubuntu/+source/eglibc/+bug/427288 | 19:10 |
| smoser | which is where we fixed the majority of it | 19:10 |
| uvirtbot | smoser: Error: Could not parse data returned by Launchpad: The read operation timed out | 19:11 |
| smoser | https://bugs.launchpad.net/ubuntu/+source/dbus/+bug/432718 has a suggestion about how you could "fix" that binary with chrpath | 19:11 |
| uvirtbot | Launchpad bug 432718 in dbus "dbus programs will not use nosegneg libraries" [Undecided,Fix released] | 19:11 |
| jiboumans | smoser: reading now | 19:12 |
| SystemTyrant | This question has nothing to do with ubuntu, but does anybody know what the shortest cat 6 patch cable can be? | 19:14 |
| === cmagina-lunch is now known as cmagina | ||
| smoser | hmm., but | 19:15 |
| smoser | $ chrpath /usr/bin/import | 19:15 |
| smoser | /usr/bin/import: no rpath or runpath tag found. | 19:15 |
| smoser | oh wait. that is a natty instance | 19:16 |
| orudie | how can i check which version of postfix server i'm using | 19:17 |
| jiboumans | smoser: if i'm reading this correctly updating dbus should fix this problem, no? if so, i have a more recent version of dbus installed than the one needed to get this patch. | 19:19 |
| smoser | no | 19:19 |
| smoser | the dbus programs were incorrectly built | 19:19 |
| smoser | and were causing that message because they would not use the nosegneg libraries | 19:20 |
| smoser | i suspect that imagemagick is similarly incorrectly built | 19:21 |
| smoser | it doesn't look like it is import, but something there is cuasing it | 19:22 |
| jiboumans | can i do somethign to help track this down | 19:23 |
| _wo__ | hey | 19:24 |
| _wo__ | hey everybody | 19:24 |
| === _wo__ is now known as _wo_ | ||
| smoser | your imagemagick is from the archive? jiboumans ? | 19:24 |
| _wo_ | i was secretly _wo_ this whole time | 19:24 |
| _wo_ | and you thought i was _wo__ | 19:24 |
| _wo_ | no need to applaud, my work speaks for itself | 19:25 |
| smoser | a simple invocation of 'import' doesnlt do it for me, and chrpath and ldd seem to idnicate import is using the correct libraries. | 19:25 |
| Daviey | We /used/ to ship a libc6-xen for this situation jiboumans ... dunno why it stopped. | 19:25 |
| smoser | Daviey, its installed in the images | 19:25 |
| Daviey | smoser, ah.. that i thought fix0rd it. | 19:26 |
| jiboumans | smoser: ii imagemagick 7:6.5.7.8-1ubuntu1.1 image manipulation programs, straight from ubuntu mirrors | 19:26 |
| orudie | any idea why I get this error when I type php -version http://paste.ubuntu.com/575126/ | 19:27 |
| smoser | jiboumans, i can't seem to reproduce it | 19:31 |
| jiboumans | smoser: my setup is xvfb-run firefox and then using import to take a screenshot of the firefox instance | 19:31 |
| uvirtbot | New bug: #728633 in squid (main) "package squid 2.7.STABLE7-1ubuntu12.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728633 | 19:31 |
| jiboumans | smoser: i'm not sure if that adds anything to the mix | 19:32 |
| smoser | i just tired running vnc4server and then inside that using import | 19:32 |
| smoser | and I dont see any messages. | 19:33 |
| smoser | what is the import command line look like ? | 19:33 |
| smoser | maybe i need to force it down a path | 19:33 |
| jiboumans | smoser: you an do it from the cli | 19:38 |
| jiboumans | let me show you | 19:38 |
| jiboumans | smoser: http://pastebin.com/RgbP1K18 | 19:39 |
| jiboumans | that's the start & screenshot script | 19:39 |
| jiboumans | you can use xvfb-run xeyes or sometthing else i'm sure | 19:40 |
| jiboumans | smoser: this is the important part for the screenshot: DISPLAY=${PORT} import -window root ${FILE} | 19:40 |
| jiboumans | default port is :99 | 19:40 |
| smoser | right. thats what i was missing | 19:41 |
| smoser | you see the libsegneg warnings in dmesg, right ? | 19:43 |
| jiboumans | yup | 19:43 |
| smoser | jiboumans, ? i dont see them here. | 19:43 |
| smoser | xvfb-run firefox http://slashdot.org | 19:43 |
| smoser | then | 19:43 |
| smoser | DISPLAY=:99 import -window root my.png | 19:43 |
| jiboumans | are you running this on a lucid 10.04.2 server ami ? | 19:43 |
| smoser | yep. | 19:44 |
| jiboumans | hum | 19:44 |
| smoser | $ dpkg-query --show imagemagick xvfb | 19:44 |
| smoser | imagemagick 7:6.5.7.8-1ubuntu1.1 | 19:44 |
| smoser | xvfb 2:1.7.6-2ubuntu7.5 | 19:44 |
| smoser | and i verified that my.png looks like a firefox screenshot | 19:45 |
| jiboumans | smoser: it only shows up once in my dmsg | 19:45 |
| jiboumans | pretty much towards the end of boot | 19:45 |
| jiboumans | the cronjob fires every few minutes though | 19:45 |
| smoser | on console ? | 19:45 |
| smoser | or in 'dmesg' output | 19:45 |
| jiboumans | 'dmesg' | 19:45 |
| smoser | yeah. i dont see it | 19:46 |
| smoser | (it probably goes to console too, just was verifying) | 19:46 |
| jiboumans | smoser: shows up in kernel log: ./kern.log:Mar 1 00:32:42 emulator-a018 kernel: [ 345.360127] ** in /lib/tls glibc libraries. The emulation is | 19:47 |
| jiboumans | that's the only time | 19:47 |
| jiboumans | which is pretty much boot time for that host | 19:47 |
| jiboumans | smoser: so we're not chasing something less important; if this is jsut slow performance, i dont really care, but my xvfb processes are crashing left and right without leaving any traces in logs, console etc | 19:48 |
| jiboumans | and i'm trying to find the root cause. this error stood out as something to investigate, but i'm not sure if it's a red herring or not | 19:49 |
| smoser | it really should just be slow performance | 19:50 |
| smoser | and i'm not really sure how slow. | 19:50 |
| jiboumans | bummer | 19:50 |
| smoser | like if its 10% or 200% | 19:50 |
| smoser | i think red herring for xvfb | 19:50 |
| smoser | but if its easily reproducible, you might just try the same xvfb in a kvm guest | 19:51 |
| smoser | and see if you can't get it to reproduce there. | 19:51 |
| jiboumans | it's reproducable on all the ec2 machines; after a whlie they just fall over | 19:51 |
| Daviey | smoser, AFAIK kvm has never had this issue... it was something xen was plagued with | 19:51 |
| jiboumans | but the issue is no logs; i'm already running it inside screen with logging turned on | 19:51 |
| smoser | Daviey, right. | 19:51 |
| smoser | it wont. | 19:51 |
| smoser | which is why i was suggesting cutting out that red herring | 19:52 |
| jiboumans | well, i'll just turn off the screenshots | 19:52 |
| jiboumans | it's a diagnostic tool i can live without | 19:52 |
| smoser | how do they go belly up ? | 19:52 |
| smoser | or fall over. | 19:52 |
| jiboumans | my monitoring tool tells me that there's no xvfb process running anymore | 19:52 |
| jiboumans | the screen.log shows firefox whinging its display went away and then crashing | 19:52 |
| jiboumans | the xvfb log says nothing | 19:53 |
| Daviey | jiboumans, What happens if you disable tls? | 19:53 |
| Daviey | jiboumans, ie, mv /lib/tls /lib/tls.disabled | 19:53 |
| jiboumans | Daviey: if it's just the screenshots creating that message, i'll just turn those off | 19:53 |
| jiboumans | i can find this out very shortly | 19:53 |
| jiboumans | smoser: syslog, daemon.log, kern.log, etc, all quiet | 19:53 |
| jiboumans | on oom killer | 19:53 |
| jiboumans | all quiet on the western front =/ | 19:54 |
| schmichael | where is a 10.04 AMI with root store as instance (not ebs)? | 19:57 |
| schmichael | these are all ebs: http://uec-images.ubuntu.com/releases/10.04/release/ | 19:57 |
| schmichael | ah, found ami-7000f019. that should work | 19:58 |
| Jonny5^ | hey guys, where can I find a .torrent for ubuntu server? | 20:00 |
| Pici | !torrents | 20:01 |
| ubottu | Maverick can be torrented from http://torrent.ubuntu.com/simple/maverick/desktop/ubuntu-10.10-desktop-i386.iso.torrent or http://torrent.ubuntu.com/simple/maverick/server/ubuntu-10.10-server-amd64.iso.torrent depending on your architecture. Other flavors can be found at http://torrent.ubuntu.com/ | 20:01 |
| Jonny5^ | thank you kind sir! | 20:01 |
| === marass_ is now known as mcas | ||
| resno | setting up fetchmail and /etc/fetchmailrc is there. what should i do? | 20:04 |
| resno | is not* | 20:04 |
| Jonny5^ | does ubuntu server include a gui for a stupid windows user like me? | 20:05 |
| resno | you can install it | 20:06 |
| resno | what are you looking to manage though? | 20:06 |
| resno | there are web based tools | 20:06 |
| Jonny5^ | I need it mainly for file sharing and chat, but it would be neat if I could use it for webbrowsing and viewing images and documents occasionally. | 20:07 |
| Jonny5^ | oh and source control | 20:07 |
| resno | surfing the web on a server? | 20:07 |
| Jonny5^ | you know, for new downloads, stuff etc | 20:07 |
| resno | or through it? | 20:07 |
| resno | im confused by your purpose | 20:08 |
| Jonny5^ | hmm. no. I actually thought of browsing. like open firefox and click stuff | 20:08 |
| Jonny5^ | as a "backup desktop" | 20:08 |
| resno | you can do that if you want, most people wouldnt recommended that | 20:09 |
| resno | you can use the desktop versino and just install the server tools you need | 20:09 |
| Jonny5^ | I know, I wouldnt recommend it myself | 20:09 |
| Jonny5^ | but that's just a minor reason. the main reason for having a gui is to be able to scroll through settings menus and stuff | 20:09 |
| Jonny5^ | you know, the way people manage windows servers | 20:10 |
| Disconnect | you won't find a lot of them to scroll through | 20:10 |
| Jonny5^ | well, I did appreciate the network settings dialog on knoppix a lot. | 20:10 |
| Jonny5^ | that was way easier than finding the proper man page | 20:11 |
| resno | you are looking for a tool like webmin then | 20:11 |
| Jonny5^ | does that come with the distro? | 20:11 |
| resno | to actually manage the machine | 20:11 |
| resno | no | 20:11 |
| resno | thats silly | 20:11 |
| Jonny5^ | :/ | 20:11 |
| resno | youll have to nistall it | 20:11 |
| Jonny5^ | what does webmin do then? is there a pm to help me install it? | 20:11 |
| resno | its not readily package with the installer | 20:11 |
| resno | youll have to do it more manual, buts what you're looking for | 20:12 |
| Jonny5^ | fair enough | 20:12 |
| Jonny5^ | I hope I don't have to build it myself though | 20:12 |
| resno | nah | 20:12 |
| resno | google for it, its harmless isntall | 20:13 |
| Jonny5^ | alright then. thanks again :) | 20:13 |
| Jonny5^ | so, risking a flame war.. why would I choose debian over ubuntu? | 20:14 |
| resno | personal preference | 20:14 |
| resno | ubuntu is debain in many respects | 20:14 |
| resno | beyond that someone else will answer | 20:14 |
| Jonny5^ | are the packages compatible? | 20:14 |
| resno | somewhat | 20:14 |
| Jonny5^ | :/ | 20:15 |
| resno | theres going to be samll differences | 20:15 |
| resno | ubuntu releases faster | 20:15 |
| resno | debain has longer rlease patterns | 20:15 |
| resno | think centos to fedora | 20:15 |
| resno | if that helps | 20:15 |
| Jonny5^ | you lost me with that last reply ^^ | 20:15 |
| resno | then forget it | 20:16 |
| Jonny5^ | k | 20:16 |
| === zul_ is now known as zul | ||
| === popey_ is now known as popey | ||
| RoAkSoAx | hggdh: ok so I'm tryting to reproduce this locally and I can't get it to work... I installed following https://help.ubuntu.com/community/UEC/CDInstall | 20:40 |
| RoAkSoAx | and the NC is not detected | 20:40 |
| RoAkSoAx | any ideas on how to resolve? | 20:40 |
| RoAkSoAx | hggdh: however, when doing euca_conf --list-nodes it does actually show the NC IP | 20:41 |
| hggdh | RoAkSoAx, look at /var/log/eucalyptus/registration.log | 20:44 |
| RoAkSoAx | hggdh: registered | 20:44 |
| RoAkSoAx | 2011-03-03 15:29:40-05:00 | 19622 -> Node 192.168.122.72 is already registered. | 20:44 |
| RoAkSoAx | [Thu Mar 3 15:44:41 2011][026938][EUCADEBUG ] DEBUG: requested URI http://192.168.122.72:8775/axis2/services/EucalyptusNC | 20:44 |
| RoAkSoAx | [Thu Mar 3 15:44:41 2011][026938][EUCADEBUG ] ncClientCall(ncDescribeResource): ppid=23816 client calling 'ncDescribeResource' | 20:44 |
| RoAkSoAx | [Thu Mar 3 15:44:41 2011][026938][EUCAERROR ] ERROR: DescribeResource() could not be invoked (check NC host, port, and credentials) | 20:45 |
| hggdh | RoAkSoAx, run 'euca-describe-availability-zones verbose', you should have a non-zero value for the limits | 20:45 |
| hggdh | huh? | 20:45 |
| RoAkSoAx | hggdh: http://pastebin.ubuntu.com/575175/ | 20:45 |
| RoAkSoAx | hggdh: this is in KVM btw | 20:45 |
| hggdh | yeap, does not seem to be there | 20:46 |
| RoAkSoAx | weird indeed | 20:46 |
| RoAkSoAx | hggdh: ahhh apparently it didn't install the -walrus :S | 20:48 |
| RoAkSoAx | package | 20:48 |
| hggdh | heh | 20:48 |
| RoAkSoAx | that might be it | 20:48 |
| hggdh | it might. But I remember Dustin has to tweak something to run fully virtual | 20:48 |
| RoAkSoAx | hggdh: yeah there's something to change from KVM to QEMU | 20:49 |
| RoAkSoAx | i just don't remnember but, regarless, it should be able to report that the NC is up | 20:49 |
| RoAkSoAx | and be able to connect to the URL | 20:49 |
| hggdh | yes, you should be able to at least see a contact | 20:49 |
| RoAkSoAx | hggdh: ill try to look little deeper to see what might be going wrong | 20:51 |
| === NightDragon is now known as GeorgeNada | ||
| RoAkSoAx | hggdh: daaaaaaaaaah dumb me... same issue as you I think :):P Didn't update to the latest eucalyptus | 21:16 |
| RoAkSoAx | lol | 21:16 |
| === Patrickdk_ is now known as PatrickDK | ||
| RoAkSoAx | Daviey: have you found out anythen else about the IFACE problem in the upstart job? | 21:24 |
| Daviey | RoAkSoAx, no | 21:25 |
| RoAkSoAx | Daviey: k that's just weird though | 21:26 |
| === NightDragon is now known as Georgenada | ||
| RoAkSoAx | Daviey: hggdh if a NC is off, should I expect to see any change sin euca-describe-availability-zones verbose ? | 21:55 |
| Daviey | RoAkSoAx, open bug about that | 21:55 |
| hggdh | RoAkSoAx, Daviey: the way I bypassed it was by taking out the instance $IFACE. The issue seem to be that this stanza is called without IFACE being set | 21:56 |
| Daviey | It is an indication NOT a promise of accuracy | 21:56 |
| Daviey | hggdh, hmm | 21:56 |
| Daviey | guess previously that was allowed | 21:56 |
| hggdh | RoAkSoAx, it should -- most of the times -- show you 0 availab ility | 21:56 |
| hggdh | Daviey, yes, it was. | 21:57 |
| RoAkSoAx | hggdh: "most" hehe the NC is off and the CLC still shows AVAILABILITYZONE|- m1.small0002 / 0002 1 192 2 | 21:57 |
| hggdh | RoAkSoAx, heh. It should update in a few, on the next contact | 21:57 |
| RoAkSoAx | hggdh: nah i'm thinking might be because of the SCHEPOLICY=POWERSAVE | 21:58 |
| hggdh | indeed | 22:00 |
| RoAkSoAx | hggdh: ok just tested... on POWERSAVE and ACTION_METHOD=0, Euca assumes that node is off and skips its checking | 22:15 |
| RoAkSoAx | hggdh: and the NC is running in PowerSave mode | 22:17 |
| hggdh | RoAkSoAx, even when the NC is *not* yet in powersave? | 22:28 |
| RoAkSoAx | hggdh: looking it simple, When the CC sends the command to Powernap, ti executes the action and goes to PowerSave... and CC assumes that node is sleeping, even tho it is not because the node is running in powersave. If something triggers any of the monitors in the NC and PowerNap detects it, it will wake up (example, if there's console activity in the NC). But the CC will have no clue of this has happened | 22:32 |
| RoAkSoAx | hggdh: however, if the CC needs to run an instance, it will send a powerwake to the NC, which will in trun recover action | 22:32 |
| hggdh | RoAkSoAx, so, even in a rather convolute way, it works | 22:33 |
| RoAkSoAx | hggdh: this of course doesn't consider "What if PowerNap goes to sleep before the CC tells it to?". This however is easly fixable by just adding a ProcessMonitor to sbin/init | 22:33 |
| RoAkSoAx | hggdh: in the perfect world, where powernap/eucalyptus timers are in sync, PowerSave works just fine, which I'm surprised but it is awesome | 22:33 |
| hggdh | yeah. One more thing to check | 22:33 |
| hggdh | RoAkSoAx, we should count our blessings... | 22:34 |
| RoAkSoAx | hggdh: After I finish my review I'll post all the info in the bug and decide what to do (either recommend to tune up PowerNap config after installation, or ship defaults) | 22:34 |
| RoAkSoAx | s/in the bug/in the bug report | 22:35 |
| hggdh | RoAkSoAx, thanks. It might be a good idea to consider it for a release note, perhaps? | 22:35 |
| RoAkSoAx | hggdh: yeah!! We need to consult this with kirkland as well | 22:36 |
| hggdh | roj | 22:36 |
| mray | is there a trick to using preseed.cfg files with a USB thumb drive? I just want it to pass the file to the installer and go | 22:58 |
| mray | I've monkeyed around with the syslinux.cfg and no variation of file or preseed/file seems to work | 22:58 |
| mray | I've read the installer guide section a few times, using the example as my preseed.cfg until I see that it's actually working | 23:00 |
| mray | thumb drive just boots into the regular installer | 23:00 |
| uvirtbot | New bug: #728752 in squid (main) "package squid 2.7.STABLE7-1ubuntu12.1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/728752 | 23:01 |
| RoAkSoAx | mray: i think you are looking for something like this: http://testcases.qa.ubuntu.com/Install/ServerWhole#Preseeded%20Installation | 23:13 |
| mray | RoAkSoAx: that's pretty close, I've seen that but I really want it to skip the first few steps as well | 23:14 |
| mray | I want an unattended reinstall | 23:15 |
| RoAkSoAx | mray: maybe cobbler? | 23:15 |
| mray | I thought preseed was the preferred way to do Ubuntu installs :) | 23:15 |
| RoAkSoAx | mray: my this? https://help.ubuntu.com/community/Installation/UnattendedCD | 23:16 |
| RoAkSoAx | mray: or http://www.debuntu.org/how-to-unattended-ubuntu-network-install | 23:18 |
| mray | I'd thought about going down that path, maybe instead of thumbdrives I'll go with PXE | 23:18 |
| mray | yeah | 23:18 |
| RoAkSoAx | mray: yeah that's probably your best option | 23:18 |
| mray | thanks | 23:19 |
| RoAkSoAx | mray: now that you are here, what's the status of zenoss. Have you guys go it working in natty? | 23:19 |
| mray | RoAkSoAx: I'm no longer with Zenoss, I'm at Opscode | 23:20 |
| mray | I'm working on Chef | 23:20 |
| RoAkSoAx | poh idn't know that | 23:20 |
| RoAkSoAx | sorry about that one :) | 23:20 |
| mray | no worries, I was recruited away | 23:20 |
| mray | and now I'm working on OpenStack :) | 23:20 |
| mray | which tends to trash server installs when it goes awry :) | 23:21 |
| lool | smoser: oy | 23:21 |
| lool | smoser: Not sure how much you care about S3 and specifically S3 + FUSE | 23:21 |
| RoAkSoAx | mray: that's cool!! and yeah I know what you mean on the awry side of things ;) | 23:22 |
| mray | I've got 3 boxes that take about an hour to reimage whenever I hose the Openstack install | 23:22 |
| lool | smoser: I had a look at the most popular implementation, s3fs, and I've sent the packaging to the Debian ITP; it worked for me albeit slowly, but I've hit a deadend: it is GPL and links to openssl with no easy way to build with gnutls :-/ | 23:22 |
| mray | so I need to automate it | 23:22 |
| lool | smoser: I've just finished packaging s3-simple-fuse, and that one is faster and much simpler to maintain (it uses boto) | 23:23 |
| mray | RoAkSoAx: gotta run, TTYL | 23:23 |
| RoAkSoAx | mray: have a good one | 23:23 |
| lool | smoser: Would you have any interest in that? | 23:23 |
| c0nv1ct | when would it be beneficial to mark a connection rather than a packet in regards to traffic shaping? | 23:23 |
| lool | smoser: If you'd like to try it out, it just finished building in my PPA | 23:26 |
| airtonix | I'm looking for a way to rename a screen in byobu without restarting the byobu session, is this possible? | 23:49 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!