hallyn_ | zul: same failures with 1.0.6. Ship it :) | 00:33 |
---|---|---|
zul | hallyn_: coolness | 00:33 |
zul | hallyn_: ill upload it tomorrow | 00:33 |
hallyn_ | i should take a look at waht debian has for qemu right now | 00:34 |
hallyn_ | 1.6.0 preview | 00:34 |
=== smb` is now known as smb | ||
=== racedo` is now known as racedo | ||
jamespage | Daviey, I'm going to add a Conflicts/Provides for the lts-raring openvswitch packages - having both versions installed at the same time does bad things | 10:10 |
Daviey | jamespage: good thinking.. not to mention the current one will fail to install anyway | 10:11 |
jamespage | Daviey, yes | 10:11 |
Daviey | The current one could also conflict with linux-lts-raring | 10:12 |
Daviey | ? | 10:12 |
Daviey | Nah, overkill. | 10:13 |
angusmcgoaway | hi all. Having a stupid issue with bind9 here. Basically i know i must be doing something wrong but can't figure out what. anyway, i added a CNAME record to a zone file contianing other (working) records.. if i do an AXFR on the dns server to list all the records then it shows up with the others but if i try to lookup the domain on the same host it returns nxdomain. | 10:36 |
angusmcgoaway | any ideas? | 10:36 |
maxb | angusmcgoaway: CNAME pointing at a name that doesn't exist? | 11:01 |
=== huats_ is now known as huats | ||
angusmcgoaway | maxb: good idea, but it does unfotunately | 11:52 |
angusmcgoaway | maxb: ahhh lol thanks! i missed the domain off the end | 11:53 |
hallyn_ | smb: remind me again, why am i looking at xen+nova (per the ubuntu server meeting)? | 12:40 |
hallyn_ | oh, am i testing it under 12.04.3? | 12:40 |
stgraber | hallyn_: does that look reasonable? http://paste.ubuntu.com/5992564/ | 12:41 |
smb | hallyn_, Not exactly. I wondered whether anybody has a way of testing my ppa:smb/xen | 12:41 |
smb | hallyn_, Some parts of nova depend on xen (well to use xen hosts). Not sure exactly how they get tied in. I believe I heard via xen-api/xcp which I could make no guarantees for atm | 12:43 |
hallyn_ | stgraber: no :) | 12:43 |
hallyn_ | stgraber: misspelled strndup toward end, | 12:43 |
hallyn_ | stgraber: is delta being leaked right now? (you add frees for it but no strndup) | 12:43 |
smb | hallyn_, So if that is the case (xcp) then testing is probably useless. Except actually to figure out whether it does work at all right now | 12:43 |
smb | (all saucy) | 12:44 |
hallyn_ | smb: so testing saucy with your ppa for nova+xen, or even for regular nova+kvm? | 12:44 |
smb | hallyn_, I got nova rebuild in my ppa with the libxen 4.3 in place, so that. | 12:45 |
smb | regular kvm+nova should be totally unaffected | 12:45 |
hallyn_ | stgraber: for the overlayfs_mount() one, I really think it'd be better to do "dup = alloca(strlen(bdev->src)+1; strcpy(dup, bdev->src)" due to how many frees you're having to add, and likelyhood of further changes resulting in a new leak | 12:45 |
hallyn_ | smb: so i should be choosing a toolstack other than xl in /etc/default/xen I assume? | 12:46 |
smb | hallyn_, With 4.3 both should work, but I guess xcp uses xl anyway | 12:46 |
hallyn_ | cool | 12:47 |
smb | hallyn_, The xl stack will use upstream qemu by default now | 12:47 |
smb | hallyn_, Just a general question. Do we have one xen host at least as part of whatever cloud testing we do? | 12:48 |
hallyn_ | i assume that doesn't mean i need to change anything | 12:48 |
hallyn_ | no idea. psivaa: ^? | 12:48 |
hallyn_ | jamespage: ^ ? | 12:48 |
jamespage | smb, not at the moment | 12:49 |
jamespage | we don't have an effective way to deploy xen using juju so its hard to fit it in | 12:49 |
jamespage | smb, zul is the test-bot for xen normally | 12:49 |
hallyn_ | stgraber: I'm sorry, I never grepped the whole source. Would you like me to convert all the remaining strdupas today? | 12:49 |
zul | jamespage: yeah too bad i dont have the hardware for it this month | 12:50 |
smb | Generally I have it only working with libvirt and xm/xl. I just want to ensure that if this gets uploaded the rest of the world does not explode | 12:51 |
hallyn_ | smb: and again by rest of the world you mean openstack on top of libvirt+xm/xl ? | 12:52 |
* hallyn_ back in awhile | 12:52 | |
smb | hallyn_, Right, as nova in theory can control xen hosts too | 12:52 |
smb | Just I have that odd memory of it not using libvirt for that (would be too easy) | 12:53 |
hallyn_ | (partitioner doesn't want to let me choose a size for the lv root... taking forever to delete lv root) | 12:53 |
stgraber | hallyn_: that'd be great, yes. (with your suggestions it looks like http://paste.ubuntu.com/5992603/, though I believe we've got a couple more in the python and lua binding) | 12:55 |
stgraber | hallyn_: (grepping for strdupa and strndupa) | 12:55 |
smb | hallyn_, IIRC nova -> xcp (open source xen-server) -> xen host. And all packages related to xcp do not compile right now in S (only exist as they got copied over). And that means slightly grumpy smb. :-P | 12:55 |
hallyn_ | stgraber: that looks much better, thanks. pls push taht with my ack, and I'll hit the rest later today | 12:57 |
stgraber | hallyn_: ok, thanks | 12:57 |
hallyn_ | hoping to test the multiple-container-start issue on btrfs today. sounds like btrfs is STILL not safe to use! | 12:57 |
setsuna_ | I have two machines 192.168.10.10 and 192.168.10.12 . The machine 192.168.10.12 is the vpn server. i have a route on 192.168.10.10 "route add -net 10.0.0.0/8 gw 192.168.10.12" which allows my servers which are on 10.0.0.0/8 to talk to 192.168.10.10. i was earlier on 10.04 and faced no issues but after upgrading to 12.04 after one or two hours one or two of the servers are not able to access 192.168.10.10 but after a network restart everything works | 12:58 |
setsuna_ | fine till another hour and again one or two random servers are not able to ping. thank you. | 12:58 |
hallyn_ | all right, bbl | 12:58 |
stgraber | hallyn_: pushed | 12:58 |
hallyn_ | thx | 13:01 |
=== ema_ is now known as ema | ||
hallyn_ | stgraber: hm, looks like ihave some spurious ERROR messaging in ppa right now... will try to address that when i do strdupas | 13:17 |
stgraber | hallyn_: damn, the new attach code broke bionic even more... | 13:33 |
stgraber | hallyn_: can you think of a replacement for those confstr calls in attach.c? appears to be missing in bionic... | 13:42 |
stgraber | (I'm almost tempted to hardcode /bin:/usr/bin:/usr/local/bin | 13:43 |
stgraber | ) | 13:43 |
hallyn_ | stgraber: do you want me to revert them for now? | 13:53 |
hallyn_ | stgraber: hm, well let's just test for confstr, and if not there hardcode the path | 13:54 |
stgraber | hallyn_: confstr is the only problem now, I solved the others (missing defines/includes), so I should just decide whether I care or I just hardcode a fallback value on bionic | 13:54 |
stgraber | hallyn_: ok, I'll do that then, easy enough | 13:54 |
hallyn_ | i doubt even opensuse cares | 13:54 |
hallyn_ | but since he went to the trouble to put it in :) | 13:54 |
spidernik84 | hi all. Any good alternative to landscape for centralized patch management? Ideally with reporting | 14:11 |
jamespage | adam_g, roaksoax, zul: https://code.launchpad.net/~james-page/ceilometer/august-fixups/+merge/180422 | 14:19 |
jamespage | adam_g, gonna test the redux charms now | 14:20 |
jamespage | adam_g, acked all your grizzly branches aside from quantum - rename of branch looked odd to me | 14:38 |
stgraber | hallyn_: oh, now it's lxc-user-nic that needs some help building on bionic ;) | 14:43 |
hallyn_ | stgraber: drat | 14:44 |
stgraber | hallyn_: missing sys/socket.h include and you're using getline without using the ifdef magic to use lxc's implementation on bionic and I've got to figure out what to do with getpwuid_r | 14:47 |
stgraber | hallyn_: ah, looks like the answer in other places was "just use the non _r version of those" | 14:47 |
hallyn_ | but that's nto thread-safe | 14:48 |
hallyn_ | i suppose we could implement our own _r functions with a pthread mutex and using local storage :( | 14:48 |
hallyn_ | mind you i'm happy to say "we ignore it on bionic for now" :) | 14:48 |
hallyn_ | just sayin' | 14:49 |
stgraber | hallyn_: do we even care about threads in lxc-user-nic? | 14:49 |
hallyn_ | lol no | 14:49 |
hallyn_ | it will have to be exec'd | 14:49 |
hallyn_ | good point | 14:49 |
hallyn_ | so yeah drop back there | 14:49 |
hallyn_ | now for the #ifdef bionic stuff, should that just go into lxc.h so everyone gets it? | 14:49 |
stgraber | hallyn_: possibly, yeah, I suppose we could regroup them in one place | 14:50 |
stgraber | hallyn_: and finally building again, sent the patchset to lxc-devel | 15:14 |
hallyn_ | stgraber: ok thx | 15:15 |
hallyn_ | zul: how do you create xen guests? xen-tools? by hand? | 15:15 |
zul | hallyn_: cloud-images should work | 15:16 |
smb | hallyn_, You might use libvirt for HVM guests | 15:16 |
hallyn_ | zul: so you'd use https://help.ubuntu.com/community/Xen "manually creating a pv guest vm" ? | 15:16 |
hallyn_ | smb: yeah but to create them | 15:17 |
smb | Or cloud-images for PV guest | 15:17 |
smb | hallyn_, HVM guest you can give an iso | 15:17 |
zul | hallyn_: more or less yes | 15:17 |
smb | hallyn_, just like kvm guest | 15:17 |
hallyn_ | ok i'll try that | 15:17 |
hallyn_ | thx | 15:17 |
hallyn_ | stgraber: hm that's a long set :) (hasn't all arrived yet) | 15:18 |
stgraber | hallyn_: yeah, 17 patches, some are just one-liners. I probably could have merged some, but I was committing as I was finding and resolving issues | 15:43 |
hallyn_ | stgraber: np, all but one acked | 15:43 |
hallyn_ | zul: I suppose it's always been this way, but seeing my dom-0 show up in 'virsh list' seems wrong :) | 15:46 |
hallyn_ | i... i.. just gotta do it: | 15:46 |
zul | hallyn_: hehe...it is | 15:47 |
hallyn_ | virsh -c xen:/// destroy 0 | 15:47 |
hallyn_ | shucks. denied | 15:47 |
smb | hallyn_, Nope when using xm thats always like that | 15:47 |
Daviey | hallyn_: I've always thought this odder... $ sudo killall init && echo wtf | 15:50 |
Daviey | wtf | 15:50 |
hallyn_ | Daviey: well, init ignores - doesn't block - the signals right? so kill succeeds | 15:51 |
hallyn_ | try -9 | 15:51 |
hallyn_ | uh, without sudo :) | 15:51 |
Daviey | $ sudo kill -9 0 | 15:51 |
Daviey | Killed | 15:51 |
Daviey | :) | 15:51 |
hallyn_ | 0 != 1 | 15:51 |
smb | down to basic wisdom now | 15:52 |
Daviey | $ sudo kill -9 1 && echo $? | 15:52 |
Daviey | 0 | 15:52 |
sgran | sure, you sent a signal successfully | 15:53 |
hallyn_ | eh, i suspect it does make sense, but am not gonna go look into it now :) | 15:53 |
sgran | did init die? | 15:53 |
=== VampiricPadraig_ is now known as VampiricPadraig | ||
sgran | just because you sent it successfully doesn't mean the application processed it the way you think | 15:54 |
Daviey | sgran: Well no, but i am saying i've always considered it a funny thing | 15:54 |
sgran | init has to be special. It is a bit funny, but it makes sense | 15:54 |
Daviey | sgran: I do know this. :) | 15:55 |
sgran | ok :) | 15:55 |
hallyn_ | stgraber: can you let me know when yor'e done pushing to git? I want to remove the two error msgs but i don't want to make you have to rebase your patches :) | 15:55 |
hallyn_ | stgraber: actually i just sent the patch to the list, can you toss it in your set? | 15:57 |
TimR | can anybody tell me tell me how to get Linksys USB100TX usb adapter to work on ubuntu 12.04.2lts | 15:57 |
stgraber | hallyn_: replied to your comment, do you see an actual problem if rand() is hit by a race with the current code? AFAICT if that happens, it'll just get seeded with some new data from urandom which should actually make it better. | 16:04 |
hallyn_ | stgraber: yeah i'm not sure, i did wonder that, but wasn't sure | 16:04 |
hallyn_ | so if rand_r doesn't exist in bionic then we can risk it. if it is then it seem sworth converting | 16:04 |
stgraber | hallyn_: I just said in my e-mail, none of the usual _r functions exist in bionic ... | 16:05 |
hallyn_ | damn android | 16:05 |
hallyn_ | k got your email | 16:06 |
stgraber | hallyn_: well, at least they solved a few stupid bugs recently | 16:08 |
hallyn_ | in their own code you mean? | 16:08 |
TimR | its showing linksys usb 100tx ethernet pegasus 0066:2203 | 16:09 |
hallyn_ | smb: well i can do 'virsh -c xen+ssh://serge@$ip list' that works, but virt-manager connecting the same way seems to hang. | 16:43 |
jamespage | Daviey, got a 'that should work' from upstream openstack on the approach to providing dkms packages for openvswitch | 16:44 |
ribo | anyone having this issue on new EC2 instances: error: unexpectedly disconnected from boot status daemon | 16:47 |
ribo | using AMI ami-11b32021 | 16:47 |
hallyn_ | smb: well... doesn't seem to be working. even manually creating a guest with the PV instructions from https://help.ubuntu.com/community/Xen fails, and when I try to start a vm i defined in libvirt I get: | 17:32 |
hallyn_ | error: POST operation failed: xend_post: error from xen daemon: (xend.err "Boot loader didn't return any data!") | 17:32 |
hallyn_ | (i did create a /usr/bin/pygrub link) | 17:32 |
hallyn_ | jamespage: hey (if you're still around) did you get a chance to straighten out iscsitarget? | 18:37 |
hallyn_ | gonna guess not based on feb changelog entry | 18:38 |
hallyn_ | ah but it's in precise-proposed, good good. | 18:42 |
bitbyte | hey guys i accidentally canceled a mv command | 18:46 |
bitbyte | is there any way to get the data back | 18:46 |
sarnold | it should all be there; the source files aren't unlinked until the destination files are closed. it might be a mess to clean up, but ought not be terrible. | 18:47 |
bitbyte | i literally canceled it like a second after | 18:48 |
bitbyte | only had chance to make 5 folders, so how would i go about cleaning it up | 18:48 |
sarnold | bitbyte: remake those directories in the source, mv those files back to their original locations | 18:49 |
sarnold | you might be able to do something like mv -i <original destination> <original source> -- but I think I'd just move files one directory at a time | 18:50 |
bitbyte | right so these were some folders holding audio files. so all i need to do is remake the parent directories and run the same command in reverse | 18:51 |
bitbyte | oh this is going to be messy :/ | 18:51 |
bitbyte | orignal command was : bitbyte@usbc01:/$ sudo mv /media/server2012/Music/* /media/sdb2/music | 18:51 |
bitbyte | and if only some of the directories were made i.e. 5 or 6 and no files copied over then i guess might be a bit took messy | 18:52 |
bitbyte | ah well iTunes will have some way of downloading it again :/ | 18:54 |
bitbyte | in the mean time any one know how to install ms-sys | 18:54 |
maxb | bitbyte: When mv is copying cross device I think it doesn't start deleting until it has copied everything. You'll want to verify, of course, but it should be trivial to check (e.g. with 'diff -rq') that there are no files in the destination which are not also still present in the source | 18:58 |
bitbyte | so if i diff -rq /media/server2012/Music i should be able to check | 18:59 |
jamespage | hallyn_, I did indeed | 19:01 |
hallyn_ | jamespage: \o/ | 19:01 |
LargePrime | Dear Server gurus. I need to ban a bunch of ips from hitting my server | 19:01 |
LargePrime | is there a tool i can just add ips to a file and they will be added to iptables as drops | 19:02 |
hallyn_ | i'd just whip up an upstart job and an add-ban script; | 19:02 |
LargePrime | =.= | 19:02 |
hallyn_ | add-ban adds the ip to a file and runs iptables to dro pit; upstart job loops over the file and drops all the ips therein | 19:03 |
LargePrime | ok so not skilled for that | 19:03 |
bitbyte | maxb: the diff -rq command keeps saying inputs wrong | 19:04 |
hallyn_ | add_ban: #!/bin/sh\necho $1 >> /etc/banned-ips\niptables -A input -s $1 -j DROP | 19:04 |
LargePrime | hallyn_: care to spoon feed me? | 19:04 |
jamespage | LargePrime, have you tried fail-to-ban? I use that on one of my servers | 19:04 |
LargePrime | adding it now | 19:05 |
LargePrime | but i not see how i can jsut paste a list of ips and get thoes banned | 19:05 |
jamespage | fail2ban rather | 19:05 |
hallyn_ | /etc/init/banips.conf:Description "ban ips"\nAuthor "me"\nstart on runlevel [2345]\n\nscript\n | 19:05 |
hallyn_ | ^ try fail2ban :) | 19:05 |
uvirtbot | hallyn_: Error: "try" is not a valid command. | 19:05 |
hallyn_ | shutup bot | 19:05 |
hallyn_ | LargePrime: fwiw you would do 'for p in ip1 ip2 ip3 ip4; do add_ban $p; done' | 19:06 |
jamespage | fail2ban also allows you to add IP's manually | 19:06 |
jamespage | as well as doing clever auth fail stuff | 19:06 |
LargePrime | i saw the auth.log autoban | 19:07 |
LargePrime | hallyn_: i need a smaller spoon | 19:07 |
hallyn_ | LargePrime: and fwow the upstart job script would just do cat /etc/hosts | while read line; do iptables -A input -s $line -j DROP; done | 19:07 |
hallyn_ | LargePrime: go with jamespage's suggestion | 19:07 |
bitbyte | any one any ideas about recovering files from a cancelled mv command ? diff -rq hasn't done anything | 19:12 |
bitbyte | and the source files haven't been copied to destination | 19:13 |
LargePrime | jamespage: how do i manualy add ips? | 19:22 |
hallyn_ | stgraber: were you going tp push your bionic-compile-fix patchset today? | 20:46 |
smb | hallyn_, fwiw PV and libvirt never worked for me. For PV guests I use cloud-images and pvgrub not pygrub (not that we ship the former) | 21:09 |
* smb crawls back under some stone | 21:10 | |
hallyn_ | smb: drat | 21:14 |
hallyn_ | and zul is gone | 21:14 |
hallyn_ | smb: but i've not gotten guests to start even with pv. | 21:14 |
hallyn_ | though yes that was with pygrub | 21:14 |
hallyn_ | (and i stopped around 1pm when the sun hit that box too hard - i'll try again tonight) | 21:15 |
jkitchen | anyone know how to make apt-cacher-ng allow certain extra things through? I'd like to use it during preseed but when apt-setup tries to grab the specified key for a repo I'm adding apt-cacher-ng won't let it through and the install fails | 21:16 |
adam_g | jamespage, those 3 issue you ran into with the python ports should be fixed now. nova-compute now ensures its OVS agent config has an IP address, and that solves the issue for me so far. | 21:36 |
Sam___ | hello | 21:44 |
Sam___ | is anyone here to give help | 21:44 |
genii | !details | 21:45 |
ubottu | Please give us full details. For example: "I have a problem with ..., I'm running Ubuntu version .... When I try to do ..., I get the following output: ..., but I expected it to do ..." | 21:45 |
jkitchen | don't ask to ask just ask | 21:45 |
twoface88 | looking for mentor ... | 22:17 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!