[08:46] <davinci11> TeTeT : Hi TeTeT, when i try to attach /dev/vdc again,  it doesnt appear in /dev, what wrong ?
[08:47] <TeTeT> davinci11: does it appear with fdisk -l?
[08:47] <TeTeT> davinci11: and what command do you use to attach?
[08:47] <davinci11> TeTeT: no
[08:48] <davinci11> TeTeT $ euca-attach-volume $vol -i <INSTANCEID> -d /dev/sdb
[08:49] <TeTeT> davinci11: there comes something to my mind - if the device is now called /dev/vdb, the complete ebs based instance stuff might not work, as the loader emi expects sdb
[08:49] <TeTeT> davinci11: attach command looks good to me, you sure it's not on the instance? does euca-describe-volumes say it is attached?
[08:51] <davinci11> TeTeT: in hybridfox show attached
[08:56] <davinci11> TeTeT: VOLUME	vol-596C061F	 1		cluster1	in-use	2011-08-03T07:37:09.288Z
[08:56] <davinci11> ATTACHMENT	vol-596C061F	i-3565075E	/dev/vdc	2011-08-03T08:41:28.731Z
[09:01] <TeTeT> davinci11: so looks good, what's on the instance to see?
[09:06] <TeTeT> davinci11: or is this the final instance with ebs attached? Then I fear it will not work due to sdb vs vdc :(
[09:06] <davinci11> TeTeT: step6  echo “1,,L,*” | sudo sfdisk /dev/vdc
[09:07] <TeTeT> davinci11: ah, you're still preparing the volume?
[09:07] <davinci11> output: /dev/vdc: No such file or directory
[09:07] <TeTeT> davinci11: fdisk -l
[09:07] <davinci11> fdisk -l
[09:08] <davinci11> TeTeT : Disk /dev/sda: 160.0 GB, 160041885696 bytes
[09:08] <davinci11> 255 heads, 63 sectors/track, 19457 cylinders
[09:08] <davinci11> Units = cylinders of 16065 * 512 = 8225280 bytes
[09:08] <davinci11> Sector size (logical/physical): 512 bytes / 512 bytes
[09:08] <davinci11> I/O size (minimum/optimal): 512 bytes / 512 bytes
[09:08] <davinci11> Disk identifier: 0x000c9641
[09:08] <davinci11>    Device Boot      Start         End      Blocks   Id  System
[09:08] <davinci11> /dev/sda1   *           1          32      248832   83  Linux
[09:08] <davinci11> Partition 1 does not end on cylinder boundary.
[09:08] <davinci11> /dev/sda2              32       19458   156039169    5  Extended
[09:08] <davinci11> /dev/sda5              32       19458   156039168   8e  Linux LVM
[09:08] <davinci11> Disk /dev/dm-0: 157.7 GB, 157659693056 bytes
[09:08] <davinci11> 255 heads, 63 sectors/track, 19167 cylinders
[09:08] <davinci11> Units = cylinders of 16065 * 512 = 8225280 bytes
[09:08] <davinci11> Sector size (logical/physical): 512 bytes / 512 bytes
[09:08] <davinci11> I/O size (minimum/optimal): 512 bytes / 512 bytes
[09:08] <davinci11> Disk identifier: 0x00000000
[09:09] <davinci11> Disk /dev/dm-0 doesn't contain a valid partition table
[09:09] <davinci11> Disk /dev/dm-1: 2122 MB, 2122317824 bytes
[09:09] <davinci11> 255 heads, 63 sectors/track, 258 cylinders
[09:09] <davinci11> Units = cylinders of 16065 * 512 = 8225280 bytes
[09:09] <davinci11> Sector size (logical/physical): 512 bytes / 512 bytes
[09:09] <davinci11> I/O size (minimum/optimal): 512 bytes / 512 bytes
[09:09] <davinci11> Disk identifier: 0x00000000
[09:09] <davinci11>      Device Boot      Start         End      Blocks   Id  System
[09:09] <davinci11> /dev/dm-1p1   *           2         258     2064352+  83  Linux
[09:09] <TeTeT> davinci11: oerks, what about paste bin?
[09:09] <TeTeT> davinci11: also this is not from an instance, it's from your client system
[09:09] <davinci11> paste bin ?
[09:10] <TeTeT> davinci11: no offense, but for getting anywhere with UEC you need to learn some basics first, I fear. Maybe visit a class on Linux/Unix shell and command line usage
[09:10] <TeTeT> davinci11: pastebin.ubuntu.com, you can copy'n'paste large volumes of text there
[09:10] <TeTeT> davinci11: so you don't spam the channel like above
[09:15] <davinci11> TeTeT: so the output ok? but yesterday when i attached /dev/vdc. it show when fdisk -l
[09:16] <TeTeT> davinci11: output is from your client system, not the instance. you need to ssh to the instance and check there
[09:16] <davinci11> TeTeT: but the same output from local machine (cc) when fdsik -l
[09:17] <davinci11> exit
[09:17] <davinci11> sorry wrong typing
[09:17] <TeTeT> davinci11: no idea what you're referring too right now
[09:21] <davinci11> TeTeT: same output from CC / using ssh from remote pc still get the same output
[09:21] <TeTeT> davinci11: no idea, I'm beat then
[09:32] <davinci11> TeTeT: Lol, i missed 1 step SSH :P can get in there and now can see /dev/vdi this time
[09:33] <TeTeT> davinci11: still you have the vdx vs sdx problem for the loader emi - no idea how to fix that
[12:14] <smoser> jamespage, ping
[12:14] <jamespage> hey smoser
[12:14] <smoser> have you, or can you, start full test of images ?
[12:15] <jamespage> I have not but I can
[12:16] <smoser> jamespage, Daviey, any one else, know of a reason that we should not test 20110802.2 ?
[12:18] <smoser> hearing none, jamespage go ahead.
[12:19] <jamespage> smoser: ack - I'll run a small cross region test first
[12:19] <jamespage> and then kickoff the big one
[12:19] <smoser> k
[13:17] <Daviey> smoser: that is our candidate.
[13:22] <jamespage> Daviey: good job - I just tested it!
[13:22] <jamespage> looks 99.95% good to me
[13:27] <Daviey> 5% is the "it doesn't boot"?
[13:27] <Daviey> jamespage:
[13:27] <jamespage> nope
[13:28] <jamespage> one test failed on the cloud-config tests - failed to mount ephemeral0
[13:28] <jamespage> but failed consistently for all instances of this test
[13:28] <Daviey> jamespage: do you test cluster compute?
[13:28] <jamespage> ah-ha - no
[13:28] <jamespage> smoser normally runs that sep.
[13:28] <Daviey> expecting https://bugs.launchpad.net/ubuntu/+source/linux/+bug/791850 to still be valid
[13:34] <smoser> jamespage, oh yeah, we're still seeing bug 784937
[13:35] <smoser> i recently fixed a t1.micro issue where the entry was written to /mnt
[13:37] <jamespage> disk is definately xdve still
[13:37] <jamespage> http://paste.ubuntu.com/657935/
[13:39] <smoser> jamespage, right.
[13:40] <jamespage> that was it tho
[13:40] <jamespage> all other tests passed
[13:48] <Daviey> So three failures?
[13:48]  * Daviey needs to prepare the release notes.
[14:40] <gtaylor> erichammond: Have you ever compared the performance of the Alestic Ubuntu AMIs to Amazon's Fedora-based golden child?
[18:35] <erichammond> gtaylor: To be clear, the recent Ubuntu AMIs listed at the top of http://alestic.com are published by Canonical, not by me.
[18:35] <erichammond> gtaylor: Comparing performance for real world applications is difficult and I have not done it.
[18:37] <gtaylor> erichammond: Ah,I was wrongly assuming that Canonical was using Alestic AMIs.
[18:38] <gtaylor> erichammond: Well, another side question. Have you had any luck getting an EC2 instance running on 2.6.39+ with the newly merged built-in Xen support?
[18:38] <erichammond> I did publish Ubuntu AMIs for two years, and apparently did a ridiculously good job branding them :)
[18:39] <gtaylor> erichammond: Yeah, your Google page ranks are still sky high
[18:39] <gtaylor> and a lot of people still reference Alestic AMIs
[18:39] <erichammond> For a while, I kept getting people who meet me say "Eric Hammond is a really familiar name."  I pointed out that they saw it every time they logged in to one of the Ubuntu instances on EC2 (back then).
[18:40] <erichammond> I expect Google to continue liking me as I still publish unique and relevant content to the subject matter being searched.
[18:41] <gtaylor> Yeah, your blog posts are interesting to see for perspective
[18:41] <erichammond> gtaylor: I don't know what you're talking about with the Xen question, but I've been using stock Ubuntu AMIs from Canonical.
[18:43] <gtaylor> erichammond: We're in a weird dead period right now. 2.6.39 no longer requires external modules for domU support (which was one of the reasons why there are specific EC2 kernels being distributed). If I understand correctly, the next releases of many distros that use 2.6.39+ can ship a much more generic kernel that can be dist-upgraded like bare metal or more tolerant VM hosts.
[18:44] <gtaylor> But almost all AMIs I see run 2.6.35 or so with the kernel either pinned or pointed at an EC2-specific package name.
[18:44] <gtaylor> IIRC, canonical uses linux-virtual or something.
[18:45] <erichammond> gtaylor: I see. I let other folks worry about that kind of thing now and I just focus at a higher level :)  You could ask smoser and the gang over on #ubuntu-cloud
[18:45] <erichammond> It's almost noon, so I should head off to work...
[18:45] <gtaylor> erichammond: OK, thanks
[18:45] <erichammond> oh, we're on #ubuntu-cloud :-)
[18:46] <gtaylor> teehee
[18:46] <gtaylor> smoser: ^^
[18:46] <erichammond> (Usually when somebody addresses me directly it's over on ##aws)
[18:46] <gtaylor> ahh
[18:47] <erichammond> later
[18:48] <smoser> gtaylor, all ubuntu amis should "just work" for upgrades
[18:48] <smoser> in 10.04, we used the '-ec2' kernel flavour.
[18:48] <smoser> in 10.10 and beyond, we use '-virtual'
[18:48] <smoser> but, mostly '-server' would have "just worked" also.
[18:49] <gtaylor> smoser: Is it safe to dist-upgrade to new distro releases on EC2? I assumed it would be for Ubuntu, at least.
[18:50] <smoser> dist-upgrade meaning 'apt-get dist-upgrade' ? right?
[18:51] <smoser> (versus 'do-release-upgrade', which brings you from 10.04 to 10.10)
[18:51] <gtaylor> well, maybe I do mean do-release-upgrade
[18:51] <smoser> it should work.
[18:52] <gtaylor> smoser: Well that's good to know, thanks
[18:53] <smoser> both should work.
[18:53] <smoser> gtaylor, if you're using instance-store, you have reason to be more worried. :)
[18:53] <smoser> but ebs can be recoverred more easily
[18:53] <gtaylor> smoser: Nope, we're using EBS
[18:53] <gtaylor> smoser: though, I have reasons to worry about EBS, too :)
[18:53] <gtaylor> (ala, the great outage)
[18:54] <smoser> this is true.
[18:54] <smoser> gtaylor, so yes, i' would suggest doing a do-release-upgrade from within a screen session
[18:54] <gtaylor> yeah, that sounds like a good idea
[18:56] <smoser> jamespage, you have link to test results ?
[18:56] <smoser> did you run the big test?
[18:57] <gtaylor> smoser: Do you have any DB servers deployed on EC2?
[18:57] <smoser> i do not.
[19:32] <smoser> jamespage, ping please.