[15:43] <paddy_m> I am having wierd issues with an ebs-backed ami running ubuntu 10.10, eventually /dev/xvdb goes to 100% cpu utilization with no actual I/O
[15:45] <paddy_m> what is /dev/xvdb it doesn't show up in fstab or "df -h "
[15:54] <kiall> paddy_m, thats just the disk, same as /dev/sda etc
[15:56] <kiall> and not showing in fstab / df -h just means it hasnt been mounted ...
[16:01] <paddy_m> how does it go to 100% i/o util with no read or write requests, and no read or write kbps
[16:03] <kiall> honestly cant tell you that, how are you sure the CPU usage is related to that disk?
[16:03] <kiall> (and not another disk.. aka a mounted one!)
[16:04] <paddy_m> cpu usage is low, iowait is high
[16:05] <paddy_m> the machine locks up, and I can't login
[16:16] <kiall> paddy_m, sorry back and forth from my desk ;) .. could it simply be a EC2 issue?
[16:17] <kiall> humm AWS status page doesnt show any issues.. probably isnt an EC2 issue then
[16:29] <paddy_m> I don't think so, this has happened on multiple instances over multiple days
[16:29] <paddy_m> we fire up the machine, it starts processing, and then after an hour or so, it ocks hard like this
[16:29] <kiall> humm .. honestly never seen anything like it before (except with EC2 issues)
[16:45] <DeaCon> just wondering what happens when mutiple user open the same file , and save at differnt times
[16:47] <kiall> DeaCon, the last one to save will overwrite the one before... not sure how thats a #ubuntu-cloud question tho ;)
[16:47] <DeaCon> so one would not get the info saved then
[16:48] <DeaCon> or can you both open files at the same time
[16:48] <kiall> exactly - just as with any normal operating system and FS ..
[16:48] <kiall> unless your talking about something cloud specific here ..
[16:52] <DeaCon> well can you get 2 systems to work on the same file at the same time
[16:52] <DeaCon> if differnt locations
[16:52] <kiall> of course...
[16:53] <DeaCon> ok i need to read more i guess
[16:54] <kiall> Honestly .. I still have no idea what your trying to do ;) Im guessing its got nothing to do with cloud services?
[16:54] <kiall> Just standard file sharing ?
[16:54] <DeaCon> im trying to open the same file and not lose info when they are saved
[16:55] <kiall> where is the file stored?
[16:55] <DeaCon> on my cloud
[16:55] <kiall> Which is .. S3? EC2? something else?
[16:55] <DeaCon> ok no idea im learning sorry
[16:56] <kiall> what do you mean on "on your cloud" then? what services / software are you using?
[16:56] <DeaCon> ubuntu one cloud
[16:57] <kiall> aha .. okay ;) Now im starting to get you!
[16:57] <DeaCon> sorry
[16:57] <kiall> (BTW lead with that sort of detail next time ;))
[16:57] <DeaCon> ok will do
[16:58] <kiall> As far as I know, Ubuntu One will, if you open the same file twice on 2x PC's, then save on PC1, then save on PC2 - PC2 changes will be saved, and PC1's lost
[16:59] <DeaCon> ok great thx
[16:59] <kiall> np
[17:01] <kiall> DeaCon, on and for UbuntuOne stuff, you'll probably get better answers in #ubuntuone ..
[17:01] <kiall> oh and*
[20:40] <cpbtklogic> Has there been a change in libvirtd since 10.04 LTS launch that makes specifying the disk type on domain definition a requirement?
[20:40] <cpbtklogic> I have XML files that used to work without problems (in libvirt on 10.04 LTS.
[20:41] <cpbtklogic> Now, I have to explicitly state the disk type... before, it used to identify that they were qcow2.  Now, everything seems to be set to 'raw'
[21:16] <flaccid> cpbtklogic: see the man page and check the package versions you have been using. if that is the case, then i assume that is the case..
[21:17] <cpbtklogic> Hi flaccid... I looked through man pages and changelog... but couldn't see anything overtly stating 'gotta specify type now'.   I'll look again.
[21:17] <cpbtklogic> If I can't find anything, I might post it to launchpad and let package maintainers figure out if it should be classified a bug or if it is a feature.
[21:17] <flaccid> well regardless, if you have to specify it, then i guess you have to.
[21:18] <flaccid> cpbtklogic: you might want to go upstream and check first, http://libvirt.org/remote.html
[21:18] <cpbtklogic> yeah... that's my take on it too flaccid.  However, if there are less aware users out there the files that used to work would now fail to define machines in really weird ways.
[21:19] <cpbtklogic> It looks to start and run fine... it is just the BIOS can't see a legit disk if you don't specify the format.
[21:20] <flaccid> cpbtklogic: you can check the upstream versions, http://libvirt.org/downloads.html
[21:20] <cpbtklogic> yes... thanks flaccid.  Will do.
[21:20] <flaccid> what package is this in, i don't see it? http://packages.ubuntu.com/search?keywords=libvirtd&searchon=names&suite=all&section=all
[21:27] <cpbtklogic> I think it is virsh.
[21:27] <cpbtklogic> libvirt-bin
[21:28] <cpbtklogic> http://changelogs.ubuntu.com/changelogs/pool/main/libv/libvirt/libvirt_0.7.5-5ubuntu27.7/changelog
[21:28] <cpbtklogic> Actually... you know what flaccid... You just made me find it.
[21:28] <cpbtklogic> "SECURITY UPDATE" ...  fix multiple issues with disk format for backing stores.
[21:30] <cpbtklogic> I think I was looking at the wrong version's changelog earlier.
[21:31] <cpbtklogic> I'm going to look up that CVE and see what it is all about.  See what it changed, that may very well explain the change in functionality I'm seeing.