[00:00] <_1094kms> ZacLnxNewb: in app channels not everyone is looking all the time - you might have to wait some time. [00:12] SpamapS: w00t! your libcgroup/libvirt waiting job seems to be working like a charm [00:58] is MX record always required for mailservers, or only if the mailserver is handling mail for another domain? [00:59] Strictly speaking, MX records are not necessary. [01:00] In practice, you will probably get significantly better results by deplying MX and SPF records. [01:23] jo-erlend: an MX record is required for servers that want to receive mail from the internet at large (the public). it tells smtp clients where to send mail for any given domain [01:24] pmatulis_: isn't the default to fall back on the A record? [01:24] twb: i never heard of that [01:24] I'm probably wrong [01:25] twb: an A record is for a host, not a domain [01:25] "If no MX records were present, the server falls back to A, that is to say, it makes a request for the A record of the same domain." [01:25] http://en.wikipedia.org/wiki/MX_record [01:25] pmatulis_: SMTP is for hosts, too [01:25] pmatulis_: this modern trend of most hosts not having MTAs is an abberancy [01:27] twb: ok, thanks for that [01:29] "Only resolvable, fully-qualified domain names (FQDNs) are permitted when domain names are used in SMTP. In other words, names that can be resolved to MX RRs or address (i.e., A or AAAA) RRs (as discussed in Section 5) are permitted, as are CNAME RRs whose targets can be resolved, in turn, to MX or address RRs." [01:29] That's 5321 (nee 2821) [01:30] jo-erlend: so short answer is: you don't have to, but we recommend you do. [01:38] hi There, I im doing backup with rsync but it stop to work in the middle apparently because long size files, somebody know how to solve or run rsync to copy big files? [01:44] twb, thanks. [01:46] I just installed Ubuntu on my desktop, using a mdadm raid and lvm on top of that. I use three 1.5GB disks and now I see /proc/mdstat tells me it'll take 22000 minutes to complete. It's running at 1MB/sec. Is that normal after an install? [01:51] 1MB, no [01:52] Not for onboard SATA ports and 3.5" 7200 drives [01:52] For external USB drives or something, those numbers wouldn't surprise me [02:38] anyone know if it is possible to browse sub directories through apache2? I have no problem seeing the files but it won't display any of the sub directories in the browser. [02:39] twb, that's what I thought. Any idea how I'll proceed to figure out the problem? [02:40] jo-erlend: dunno, iostat, smartctl short self-tests, swap in a new SATA cable, etc.? [02:40] thesheff17: have you turned on directory indexing? [02:40] twb, oh.. Actually, I did notice that one of the sata cables were looking a little strange. [02:41] twb, do you think a faulty cable could slow things down without causing other problems? [02:43] twb: checking now [02:45] jo-erlend: it's easy to swap in a new one, so do so [02:45] jo-erlend: also recommend turning on write-intent bitmaps, so that you don't need to restart the sync every time you reboot [02:45] Hi there I extended a logical volume using lvextend, when I do lvdisplay the new size look correctly, but when I do df -h still show me the old size. There is something that I need to do after/before lvextend? [02:48] twb: DirectoryIndex looks like what apache2 defaults to when it his that directory like index.html. Am I looking for another param? [02:48] Roxyhart0: you have extended the size of the logical volume, and now you need to extend the size of your filesystem to match that additional space. [02:48] RoxyHart0: It's as if you copied a partition to a larger hard drive without expanding it. [02:49] Roxyhart0: assuming you're using ext2, ext3, or ext4 check out resize2fs [02:49] :s/his/hits [02:50] ok, i understand ...yes I am using ext3 [02:50] its not dangerous? [02:52] Roxyhart0: I've done it, on my personal data, without a backup when I was feeling brave. If this is mission critical, go buy a cheap multi-Terabyte hard drive to backup the data on. [02:55] yes, thanks so much...I am backing up again and then will do the resize [02:55] Roxyhart0: good luck! [03:27] jjohansen, did you get anything sorted out with zul on how to get a haryd kernel? [03:39] twb: IndexOptions ShowForbidden jeeze this was it....found it in a really old forum...I guess it was because I was securing sub directories with .htaccess files. [03:40] twb: w/ different username/pass [03:40] That reminds me, I need to disable .htaccess on my apache2 [04:04] smoser: afaik still looking [04:07] RoAkSoAx: uploaded [04:18] zul thanks!! === airtonix_ is now known as airtonix [04:43] hi somebody know how I can set a linux server to send email but as client ? [04:50] hello guys [04:50] i'm on a dialup link outside work [04:50] i'm using openvpn [04:51] i'm having huge packet loss over this link, on my ISP's route it's almost no packet loss, whenver I turn the VPN in, then I've got this intermediate packet loss on different intervals and different times, does anybody at #ubuntu-server have a clue to point me in the right direction, if any, to fix this? [04:53] anyone???? [05:10] mtu? [05:14] hello, i have an ec2 vs ubuntu server question if anybody's got any time [05:49] willkessler: ask away [05:50] thesheff17: i've been getting some help in #ubuntu-cloud which might be more appropriate... although they basically just said switch to maverick [05:50] I've got an issue with my sudoers file [05:51] thesheff17: i could ask you one thing, what's the best maverick AMI for ec2 as far as high-load web server starting point [05:52] Today i logged into my server (which only has internal access and single user, me) to the message "iarp not in sudoers file ....". My old byobu session was still running to i connected to that and i can sudo all i want in here without errors, but now i'm worried if i close this session i will lose all sudo abilities, ever heard of this issue? [05:52] maverick: why maverick vs lucid? http://uec-images.ubuntu.com/releases/10.04/release/ here are the ami for lucid [05:54] flaccid suggested switching to maverick because of load issues my app is experiencing at random times, whcih do not correlate to app usage at all [05:54] however, i just read this comment that gives me great pause: http://alestic.com/2010/10/ec2-ubuntu-maverick#comment-484 [05:55] allright, i might be forced to switch to centos... sigh! [05:56] willkessler: are you experiencing problems with ubuntu in the cloud already with lucid? [05:56] thesheff17: correct, i am [05:57] willkessler: hmmm...I run a couple web sites in the cloud but I guess I don't see the volume you are getting. [05:57] willkessler: what is happening? during high load? [05:58] thesheff17: we have volume of about 3500 new users on a social game, per hour. between 3-6pm the cpu will spike from about 3 to 40-60 [05:58] however, the number of users may only increase 10-20% [05:59] nothing is changing at the code level; the game does not write anything to the ebs-backed drives on the instance [06:00] the app only talks to mysql and redis, but neither reports any major load increase [06:00] so i am thinking, something about higher network traffic spikes a big cpu usage increase [06:01] willkessler: is it apache2 that is spiking the load? [06:02] thesheff17: yes, it appears to be. so normally ifi were listening to myself here i'd say, fix your own )*@)#*@ app [06:02] but in this case, i don't think it's the app per se because I/O to the box isn't changing enough [06:02] (acording to iostat) [06:03] so i'm thinking, wow, something about 10.04 on ec2 has varying load during times of the day, all by itself [06:03] willkessler: are you tuning apache? [06:03] thesheff17: we have heavily tuned it, yes [06:04] thesheff17: for example, we have a fair # of spares, but not so many to threaten to swap; we have carefully tuned keepalives to match the apps needs... and the thing is very consistent during most of the day [06:04] thesheff17: only during this "peak" time does it explode, but as i say, not because 30x the usage is happening [06:05] willkessler: it does sound like an I/O issue maybe to the backend db.... [06:06] thesheff17: a decent hypothesis, yes, although RDS does not report much by way of heavier activity either [06:06] ok i'll keep digging, thanks for your help! [06:07] Whats the default groups for the first user on a fresh install? [06:07] willkessler: well if the pipe is full on the network...then the db won't show higher load [06:08] iarp: what do you mean on a fresh install? during the install you usually pick the username/password. [06:10] thesheff17: i've somehow lost the ability to "sudo" outside of this byobu session. [06:10] iarp: you should be able to ssh to root. [06:11] thesheff17: what's roots default password? [06:11] that is also set during the install of ubuntu server? [06:11] how did you install it? [06:12] it's been perfectly fine for months now until i logged in today, i can't remember anything i've changed for this to happen [06:14] i just don't know what groups the first user/admin is in because i may have screwed that up, i'm currently in www-data and my own usernames group, are admins in any other groups that would grant sudo access [06:15] willkessler: https://www.cloudkick.com/blog/2010/jan/12/visual-ec2-latency/ not good :-/ [06:17] iarp: well that isn't good....you may be able to find out the default username by see the /etc/passwd file. Do you also have physical access to the server? You could recover the root password easily with a live cd. [06:17] hi there, somebosy know how I can set sstmp using by default? [06:18] thesheff17: i just started up a virtualbox idental to the server install, i've somehow removed the groups "adm, dialout, cdrom, plugdev, lpadmin, sambashare, admin" from my user. I'm guessing that happened when i was having permission issues last week. [06:19] iarp: yea the default user during install is part of the adm group. [06:20] iarp: the admin group is the one that determines sudo use, by default. Is any user at all in the admin group? If so, log in as that user and sudo should then work [06:20] thesheff17: admin not adm [06:20] currently my only groups are "iarp, www-data" [06:21] i was having issues last week with apache not being are to write to directorys and i must've forgotten -a on my usermod's -G [06:21] iarp: Then your current user cannot run sudo. So, check what otherusers on the machine exist and which ones are in the admin group. [06:23] iarp: For example, try grep ^admin /etc/group [06:23] "sudo usermod -a -G adm,dialout,cdrom,plugdev,lpadmin,sambashare,admin iarp" and we're back in business, thanks. [06:24] grep ^admin /etc/group - "admin:x:111:iarp,www-data" [06:25] Why is www-data in the admin group? I'd rate that a security risk. [06:25] i'm not sure, it's being removed [06:26] willkessler: a couple more things to check is during high load what is the latency?...are all the ec2 instances in the same zone? Just for testing maybe trying running the database on localhost so the apache2 doesn't have to go out to the network. Build a layer in front of apache2 like squid to reduce load on apache2. [06:27] iarp: OK. Now you are "safe", you might also want to consider adding a user iarpbackup who is in the admin group and who has a password you will remember... [06:29] Roxyhart0: I would read here: https://help.ubuntu.com/community/Postfix [06:30] jmarsden: Well now that i know what i've gone and done i'll be extra careful next time i mess around with adding/removing groups [06:32] iarp: Yea I rarely have to mess with the default groups ubuntu comes with. If you need apache to write to a dir just do chown -R www-data:www-data /pathToDir/ assuming apache2 is running under www-data. [07:17] hello everybody [07:18] im back as promised [07:18] hello ikonia [07:18] hello jamespage [07:18] how r ya'll tonight? [07:25] so can ya'll help me with my problem tonight? [07:25] i updated the kernel last night to the newest ver [07:27] and still having the wvdial.conf pppd option problem along with the /dev/ttyUSB0 device busy problem [07:31] New bug: #721108 in exim4 (main) "Regression in 4.71-3ubuntu1.1 (lucid) when using macros" [Undecided,New] https://launchpad.net/bugs/721108 [08:03] Anyone able to assist debugging a networking issues on a new server? [08:15] anyone here? [08:15] yes [08:15] can u help me fix this http://pastebin.com/DL92R0Vp [08:16] MadHaTTer_777: haven't touched PPP in ages, sorry. Where is this log from? pppd? [08:17] yes [08:17] yeah it seems like no one has used pppd or wvdail in forever thats what everyone tells me [08:17] MadHaTTer_777: is /dev/pts/5 a modem for real? [08:18] pts/5? [08:18] should be /ttyusb0 [08:18] thanx for pointing that out tetet [08:19] do u know how i change that by chance? [08:19] MadHaTTer_777: np, fear I'm not that helpful with such a problem though [08:19] k [08:19] MadHaTTer_777: guess 'pppd /dev/tty/usb0' or whatever the correct one is [08:21] k ill try [08:22] nope [08:36] MadHaTTer_777: have you logged the bug as suggested [08:36] MadHaTTer_777: have you done the testing work as suggested [08:37] MadHaTTer_777: have you detailed if the modem is supported by Linux, if so what versions/kernel/dependencies ? [08:41] What could cause all http connection to not connect? [08:41] I've setup a server and ping, and dns work [08:41] but http connections fail [08:42] as well as git clones fail to connect [08:43] what testing work was suggested and ill tell u if i did it [08:43] i dont log bugs until i am sure they are bugs and not just pebcak [08:43] and the anwser to ur 3rd q is yes and yes [08:52] MadHaTTer_777:1.) confirm the modem is supported under Linux 2.) confirm it's dependencies/version requirements 3.) if 1 and 2 are positive test on the desktop kernel 4.) report bug if 1+2 are met and 3 fails [08:54] quentusrex: check the processes are running, check the routing, check the firewall rules, that's a good starting point [08:56] u caught all im looking for is how to enter a little code in wvdial-pipe right? and u want me to go through all that hell to figure it out? [08:57] btw most of that cant be done anyways [08:57] ikonia, I know something works because the kvm guests have working network connection [08:58] MadHaTTer_777: ok you have a choice, write the code yourself, or go through the bug reporting process [08:58] quentusrex: ahhhh kvm....ok, so that means your routing for the host rather than guests maybe off [08:59] ikonia, probably [08:59] quentusrex: first few times I setup kvm I fell into some stupid routing traps with host/guest networking getting quite mixed up [08:59] I still fall into some now, even with more experience [09:00] ok ikonia the code im talking about is probally 2 words long [09:00] so ? [09:01] how does that change anything ? [09:02] its probally basic linux commands like pppd = /dev/ttyusb0 or something like that but not knowing linux to the extreme gives me a disadvantage which is why i am trying to ask someone who would know [09:03] MadHaTTer_777: have you checked the device is supported in linux, yes/no ? [09:04] yes [09:04] ok, what are it's support requirements (Linux/kernel version) [09:05] ubuntu the last known ver was 7.4 because the manufactor hasnt updated their site since before 7.4 came out [09:06] MadHaTTer_777: can you show me the link to the manufacture page please ? [09:06] (or tell me the device named/make/model and I'll look it up) [09:08] franklin wireless mobile modem cdu-680 p/n blah blah 0014 [09:08] carrier ACS manufactor Franklin Wireless [09:08] rmoning [09:09] MadHaTTer_777: ok, I'm reading there is a 3rd party module needed for it, have you installed that MadHaTTer_777 ? [09:09] ? [09:10] where are you reading that? [09:10] on the FAQ page [09:10] http://www.franklinwireless.com/?doc=bbs/gnuboard.php&bo_table=faq1&wr_id=2 [09:12] im not seeing anything about a 3rd party module there [09:12] software installed, it's in the "Linux" folder [09:12] under "what software should I install" [09:13] OS support for that 680 is Windows [09:13] http://www.franklinwireless.com/?doc=bbs/gnuboard.php&bo_table=product&wr_id=4 [09:13] ohh that lol [09:13] so lets see, it says it supports Windows, not linux [09:13] yeah that software is garbage [09:14] well, there you go then [09:14] so you need to contact the vendor [09:14] no ikona [09:14] yes [09:14] http://www.franklinwireless.com/?doc=bbs/gnuboard.php&bo_table=product&wr_id=4 [09:14] have u not listened to anything ive said over the last few days? [09:14] os support Windows [09:14] I have [09:15] There Linux support is nonsense they say they support ubuntu 7.1 and are testing 7.2 [09:15] yeah, sounds like nonsense [09:15] therefore anything beyond 7.1 (??? 7.l0 maybe) is untested [09:15] their software is crap so i compiled one that works much better from a script someone wrote and posted in the ubuntu forums [09:16] clearly it doesn't work much better as you're saying it's not working in here [09:16] its not that that is having the problem [09:16] its some of the linux files it uses [09:17] as I understood it you couldn't find the right device file for it ? [09:17] i have properly tweaked and compiled that and rewritten the wvdial.conf and widial.conf but the pppd still needs to be tweaked so im looking for how to go about tweaking it [09:18] you don't compile wvdial.conf, so what are you talking about [09:18] lol [09:18] ik [09:18] and what is "the pppd" [09:18] if you know, why are you saying it ? [09:18] I'm not laughing, if you want help, communicate properly with information rather than random words such as compiled wvdialconf [09:19] compile was referring to the code for the device that i forgot to mention in that sentance [09:20] and its wvdial.conf [09:20] ;P [09:20] MadHaTTer_777: right, where did you get the "code" for this device and is it a kernel module you built ? [09:21] just a sec lemme find the link [09:21] no problem [09:26] http://ubuntuforums.org/showthread.php?t=682596 [09:27] the code is at the bottom [09:28] MadHaTTer_777: when you built this code (I'm reading it now) what did it build ? [09:31] well its a usb flip flop device between scsi flashdrive and ppp connection over port 110 on a cell band so this made it so i could switch between them then i rewrote as described in later pages on that post the wvdial.conf and such [09:31] MadHaTTer_777: no, what did it actually build [09:31] as it what is the name of the file it built [09:32] i think it was changemode [09:32] just a sec ill look [09:32] yeah changemode [09:34] oh and their old franklin software also used wvdial acording to the execute.sh [09:34] forget wvdial for a minute [09:34] k [09:35] as I'm reading this the changemode binary swaps the device between a mass storage device and a modem device. Which makes it loaded the required kernel modules (and therefore create a device file) [09:36] when you put the device into "modem" mode, from what I read it does a check to see if the device is mounted (is your device mouted ?) then creates a device file for the modem, has that device file been created ? [09:37] yeah like i said a scsi and ppp over cell band - no its unmounted [09:37] yes [09:38] ok, what is the name of the device file that is being created [09:38] sprintconfig [09:38] (for the modem) [09:39] that's not the name of the device file, what's the name of the device file it created in /dev [09:39] ohh that [09:40] as i recall ttyusb0? [09:40] ok as you recall, you need to be certain of that [09:41] MadHaTTer_777: reboot your machine with the device is "disk" mode, ls -la /dev > devicelist.out then change the device to a modem mode, then ls -la /dev > new_devicelist.out then diff the two === fceccon is now known as fceccon|a [10:03] How do I know what lv is /dev/dm-4 ? [10:05] hey ikona i just checked something and its actually on ttyusb2 [10:06] actually these readouts r kinda strange just a sec and ill send [10:11] hmm [10:20] gonna be hey ikon? [10:21] -gonna be [10:21] +ia [10:21] How do I know if my hardware has IPMI support? [10:22] lshw? internet? [10:24] Or I guess just install ipmitool and see what it says [10:25] hey ikonia see the reason im pretty sure its not the modem is it connects then when using the ppp0 interface it picks up all the dns addys then just kinda sits there [10:25] yeah im sure trial and error would work to [10:47] <_ruben> twb: modprobe the various ipmi kmods and check dmesg? [10:48] _ruben: I ran sensors-detect for something else, and it apparently checked for IPMI stuff for me [10:48] Anybody know offhand if/what cpufreq module AMD Geodes use? [11:02] Is there any way to see why my server takes so long to load pages? [11:02] When local it takes a longer than normal amount of time to load [11:34] MadHaTTer_777: sorry, had to step away for a minute [11:34] MadHaTTer_777: if the thing connects and gets an IP, why are you using wvdial ? [11:35] it connects using wvdial [11:35] but hangs [11:35] MadHaTTer_777: what IP does it give you ? [11:35] after the dns adresses [11:39] i dont even know how to check that in linux [11:39] New bug: #721207 in openssh (main) "Dependance problem install openssh server" [Undecided,New] https://launchpad.net/bugs/721207 [11:39] whats the line? [11:40] MadHaTTer_777: can you pastebin the output of "ifconifg -a" please. [11:46] i did [11:46] i did [11:46] i did [11:46] wjefoad [11:46] opps my bad didnt mean to flood [11:47] i pm'ed it 2 u ikonia [11:51] did u get it ikonia? [11:52] sorry, was away again [11:52] ah, I see it, one moment [11:53] MadHaTTer_777: ok, I think we are getting somewhere, can you pastebin the output of the following 2 commands please "cat /etc/resolv.conf" and then netstat -rn [11:53] k didnt really feel like sharing ip's with everyone [11:53] k [11:53] no, I fully understand that [11:57] http://pastebin.com/K3P9WAK0 [11:58] MadHaTTer_777: ok, so based on the IP you gave me I can see why nothing is working [11:58] ok why? [11:59] MadHaTTer_777: if you look at that routing table, everything (0.0.0.0) is going out of 192.168.8.1 which is your ip on your eth0 card [11:59] ok how do i change it in linux? [11:59] ohh wait [11:59] no thats not it :( duh [12:00] ? [12:00] what's not it ? [12:00] hallyn, Is http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=604735 likely to be an issue for us? [12:00] i currently have both hooked up how do u think im pastebin... [12:00] Debian bug 604735 in seabios "new version of seabios resolves "guest moved used index from ..." when using virtio-blk" [Important,Fixed] [12:00] hallyn, (I haven't attempted to reproduce it) [12:00] MadHaTTer_777: ok, what are you expecting to go out of the ppp interface ? [12:01] well nothing i wanna switch it over [12:01] but at the moment thats how i have it hooked [12:01] ok, at the moment then all your traffic is going out of eth0, that's why [12:02] you'll have to switch it so your default gateway is the ppp0 device for all traffic to go over it [12:02] ok how do i do that ? [12:02] eth0 unplugged now [12:03] well, the simple thing would be to disable your eth0 network, reboot and then ppp0 will become the default route when you start it up, other than that you can use "route add" and "route delete" to manipulate the routing table [12:03] (by disable I mean disable the interface, not just unplug it) [12:03] i think ill go with the former and have a smoke with the wife while it boots [12:04] brb [12:12] nope didnt work wife wants me to come to bed will u be here tomarrow? [12:58] Daviey: no, we have qemu 0.12.5 in maverick, and natty has newer seabios [12:58] so shouldn't be [13:00] (and actuallly we have newer seabios in maverick) [13:08] hallyn, oh good.. thanks. [13:26] Anybody know much about writing custom collectd data snarfers? [13:27] I'm trying to teach collectd to graph the SNR on my ATM card by catting values from /sys/ [13:38] hallyn, What was the status with the multipath merge? You did tell me, but i can't remember what you said.. sorry. [13:57] morning all [14:01] OK, this is fucked up. [14:01] I have unqualified hostnames in /etc/hosts. [14:02] The host itself has 8 interfaces, so there are eight lines like "a.b.c.d alpha". [14:03] If I change *ONE* of those to "a.b.c.d alpha.cybersource.com.au alpha", then 1) "hostname -f" works; and 2) "getent hosts alpha" gives output like this: http://paste.debian.net/108062/ [14:03] WTFingF, man [14:03] (And why can't I just bloody run "dnsdomainname cybersource.com.au" at boot or something...) [14:10] Daviey: status was cmagina had done some testing and found issues during FOFB, and asked me to look at the logs [14:10] cmagina: i did take a peek but saw nothing that i've not seen with the current version of multipath :) [14:11] cmagina: does current mp do better than the debian/experimental merged version? [14:11] Daviey: i might have to try the debian/unstable version... [14:12] hallyn: using the experimental version, the machine was constantly failing over while idle. The current version we are using, 0.4.8, everything comes up ok [14:13] cmagina: oh, so you'd never done a failover in those logs? [14:13] hallyn: the experimental mp was a mess, causing IO errors and scsi eh to kick in, while the system was just idling [14:13] hallyn: nope [14:14] hallyn: both SCMs were powered on [14:14] jinkeys [14:14] hallyn: all that stuff was just from booting and idling after boot [14:14] yeah, not pretty [14:15] well trying an unstable merge is gonna be pretty low on my list the next few days [14:15] Daviey: did you ask bc you are needing it for something? [14:15] or bc feature freeze is coming up, i guess? [14:16] bc gets new features? [14:17] I thought it was just a calculator [14:17] Oh, you meant ∵. [14:19] :) [14:22] If you were IRCing from emacs you'd have bitchin' UIMs like TeX and SGML as well as boring ol' greek and katakana and such. [14:22] hallyn: do you want a debdiff for libvirt lxc fixes that i cherry picked from upstream or bzr branch? [14:24] zul: whichever is easier for you [14:24] zul: if they're cherrypicked from upstream, then maybe we should just take a new snapshot? [14:24] BTW, I fixed my solos vs. collectd fight [14:25] hallyn: its just one right now but yeah there is a newer version of libvirt available but i dont know how that is going to set you back [14:25] zul: good point, i don't know thta i have time for that [14:26] hggdh: bug #711587 [14:26] Launchpad bug 711587 in eucalyptus "powernap and Eucalyptus seem unable to reach an understanding" [High,New] https://launchpad.net/bugs/711587 [14:26] hallyn: since i had like 3 lxc libvirt instances running and been consoling in and out all day yesterday and didnt have a problem with the patch i cherry picked [14:26] awesome [14:26] so im going back to libvirt ;) [14:26] hm, i was looking at several bugs, but i suppose i should be prioritizing merges right now? [14:26] hggdh: so what's the real deal there? :) [14:27] hallyn: depends on the merge [14:27] libvirt and multipath, at least [14:28] hallyn: if it fixes alot of bugs then i would do the merge and kill two birds with one stone [14:28] but thats just an opinon ;) [14:29] http://alioth.debian.org/~twb-guest/internode.png and extel.png [14:31] zul: no, the bugs are largely in older releases [14:32] hm, and then i was putting off the cap_syslog manage update, but i suppose it would be nice to get that into natty [14:34] RoAkSoAx: I am assuming this is a fluke; I will retest, with the new euca Daviey has uploaded [14:34] hggdh: I'm assuming this is powernap issue :) [14:35] RoAkSoAx: i think you might have killed "powernap now" support [14:35] RoAkSoAx: apt-get source eucalyptus [14:35] RoAkSoAx: grep -r powernap [14:35] hggdh: so, nodes go idle. powernap puts them to sleep when it shouldn/t? [14:35] RoAkSoAx: let me test it first with the new source [14:35] RoAkSoAx: and you should see how eucalyptus calls powernap [14:35] RoAkSoAx: eucalyptus does not use the power daemon at all [14:35] RoAkSoAx: it triggers its own sleeps and wakes [14:36] kirkland: yeah I saw it, :) powernap-now works. [14:36] RoAkSoAx: what I had was nodes going on powersave and euca losig reference to them -- not bringing them back up [14:36] hggdh: show me powernap's config [14:36] RoAkSoAx: but this euca was a bit unstable [14:36] oh goody [14:36] Daviey: this is before ypur 2.0.1 [14:37] RoAkSoAx: no chance now, machines have been reconfigured for Lucid [14:37] kirkland: I think the issue is related to either powersave or that it monitors for other activity than rather kvm process (such as load and stuff) [14:37] hggdh, hmm.. i'm not confident it'll be fixed [14:37] Daviey: neither am I, but I am also not confident this is not a red herring [14:38] hggdh: will a 2 node setup in KVM allow me to test the issue? [14:38] as soon as I finish the SRU kernel tests I will reconfigure the machines for natty [14:38] RoAkSoAx: it should [14:39] hggdh, btw, would you mind leaving the machines running clean natty... would like to do some testing monday morning, whislt you are sleeping? [14:39] (assuming you don't need them over the weekend) [14:40] hggdh: ok, I'll try to setup it up and see if I can reproduce. The thing is that in powersave mode, the machines keep running, so euca might think they are still alive or something like that [14:42] RoAkSoAx: it might. I am sorry, but this week was hell (and the previous) [14:43] Daviey: no prob. Just be careful not to step on zul, he is using cempedak [14:43] i should be ok Monday when im sleeping [14:43] RoAkSoAx: do you have access to these machines? [14:44] hggdh: no worries! :). I just wanna make sure this gets fixed so you nor me have any headaches later :) [14:44] hggdh: uhmmm I don't know actually [14:44] hggdh, wilco [14:46] New bug: #721295 in apache2 (main) "After setting up apache2 ssl error "ssl_error_rx_record_too_long"" [Undecided,New] https://launchpad.net/bugs/721295 [14:52] Daviey: you want 2.6.38-4 or -3 as kernel? [14:57] smoser, does https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/656646 still exist? [14:57] Launchpad bug 656646 in cloud-init "transient network/keyscan issues" [Low,Confirmed] [14:58] skaet_, umm.. i cant say it doesnt. but i also can't say that it wasnt transient network issues. [15:00]