| daxt | Hello | 00:22 |
|---|---|---|
| daxt | i get this error "Looking up studiopaceatl.com | 00:22 |
| daxt | Cannot look up hostname studiopaceatl.com: Name or service not known" in ubuntu EC2 instance | 00:22 |
| daxt | but ftp server works and i am connected to using gftp | 00:23 |
| daxt | can somebody direct me to the right path ? | 00:23 |
| daxt | anyone home ? | 00:25 |
| daxt_ | HHHHHHHHHHEEEEEEEEEEEEEEEEEEEELLLLLLLLLLLLLLLLLLLLLOOOOOOOOOOOOOOOOOOOOOOOOOWWWWWWWWWWWWWWWWWWWWWWWWWWW | 00:36 |
| erichammond | daxt_: From what you've described so far, this does not sound like an issue that would be specific to EC2. | 00:44 |
| daxt_ | hi erichammond , then what it can be ? | 00:45 |
| erichammond | Note that it sometimes takes hours for EC2 and UEC questions to be answered on this channel. Non-cloud questions will probably get better help elsewhere. | 00:45 |
| daxt_ | so mine is non-cloud ? :D | 00:46 |
| daxt_ | can u point me to a proper place which might help me to resolve this | 00:46 |
| erichammond | daxt_: It sounds like a DNS lookup issue for a general host name. DNS for this on EC2 works pretty much the same as anywhere else. | 00:47 |
| daxt_ | actually my ftp gui client connects and works properly | 00:47 |
| daxt_ | only the ftp cli at ec2 instance does not manage to lookup the host , is the problem | 00:47 |
| daxt_ | shall i modify the resolv.conf and add externam dns than what amazon gives ? will it help me ? | 00:48 |
| erichammond | daxt_: Sounds like you'll need to investigate how the name is being resolved for your local client and how your local environment differs from the environment on your server. I don't have enough information about any of that to be able to help you sufficiently | 00:49 |
| daxt_ | erichammond , when i connect to the ftp hoststudiopaceatl.com from my PC , it connects , but when i try to connect from the EC2 instance using ssh , it does not | 00:50 |
| daxt_ | so it has to be a dns issue on EC2 according to my feeling | 00:50 |
| erichammond | daxt_: I'm not telling you to go away, just pointing out that it probably isn't directly related to the topic of the channel, and given the nature of this channel, you might find better help over on #ubuntu or other places. The fact that it is not working on an EC2 instance is unlikely to be related to EC2 as a service. | 00:50 |
| daxt_ | erichammond , let me show u a screenshot , so you can get a clear understanding :) | 00:54 |
| daxt_ | http://img820.imageshack.us/img820/8213/screenshot18f.png | 00:56 |
| daxt_ | this issue is EC2 related | 00:59 |
| daxt_ | EricHammond , could this be an issue with security groups ? | 01:00 |
| daxt_ | aka amazon firewall ? | 01:00 |
| === daker_ is now known as daker | ||
| waltc | I realize this is semi vague here (detail in forum) but is it reasonable to get warnings and or errors during a startup of UEC 2 (10.10)? | 16:37 |
| waltc | the first one is | 16:39 |
| waltc | The file cloud-error.log contains the following: | 16:39 |
| waltc | 10:54:16 ERROR [SystemUtil:Thread-23] com.eucalyptus.util.ExecutionException: ///usr/lib/eucalyptus/euca_rootwrap losetup /dev/loop0 error: loop: can't get info on device /dev/loop0: No such device or address | 16:39 |
| waltc | the second one is: 10:54:17 ERROR [ISCSIManager:Thread-23] com.eucalyptus.util.EucalyptusCloudException: tgtadm: invalid request | 16:42 |
| waltc | Since I don't have iscsi implemented that may explain this but why there error? | 16:43 |
| waltc | the third one is:10:54:20 ERROR [NioServerHandler:New I/O server worker #1-1] Internal Error. | 16:48 |
| waltc | com.eucalyptus.ws.ServiceNotReadyException: System has not yet completed booting. | 16:48 |
| waltc | at com.eucalyptus.ws.server.NioServerHandler.messageR eceived(NioServerHandler.java:107) | 16:48 |
| waltc | at org.jboss.netty.handler.stream.ChunkedWriteHandler .handleUpstream(ChunkedWriteHandler.java:114) | 16:48 |
| waltc | at org.jboss.netty.channel.Channels.fireMessageReceiv ed(Channels.java:385) | 16:48 |
| waltc | at org.jboss.netty.handler.codec.replay.ReplayingDeco der.unfoldAndfireMessageReceived(ReplayingDecoder. java:459) | 16:48 |
| waltc | at org.jboss.netty.handler.codec.replay.ReplayingDeco der.callDecode(ReplayingDecoder.java:443) | 16:49 |
| waltc | at org.jboss.netty.handler.codec.replay.ReplayingDeco der.messageReceived(ReplayingDecoder.java:381) | 16:49 |
| waltc | at com.eucalyptus.ws.handlers.http.NioSslHandler.hand leUpstream(NioSslHandler.java:31) | 16:49 |
| waltc | at org.jboss.netty.channel.Channels.fireMessageReceiv ed(Channels.java:342) | 16:49 |
| waltc | at org.jboss.netty.channel.Channels.fireMessageReceiv ed(Channels.java:329) | 16:49 |
| waltc | at org.jboss.netty.channel.socket.nio.NioWorker.read( NioWorker.java:330) | 16:49 |
| waltc | at org.jboss.netty.channel.socket.nio.NioWorker.proce ssSelectedKeys(NioWorker.java:282) | 16:49 |
| waltc | at org.jboss.netty.channel.socket.nio.NioWorker.run(N ioWorker.java:203) | 16:49 |
| waltc | at org.jboss.netty.util.internal.IoWorkerRunnable.run (IoWorkerRunnable.java:53) | 16:49 |
| waltc | at org.jboss.netty.handler.execution.MemoryAwareThrea dPoolExecutor$MemoryAwareRunnable.run(MemoryAwareT hreadPoolExecutor.java:502) | 16:49 |
| waltc | at java.util.concurrent.ThreadPoolExecutor.runWorker( ThreadPoolExecutor.java:1110) | 16:49 |
| waltc | at java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:603) | 16:49 |
| waltc | at java.lang.Thread.run(Thread.java:636) | 16:49 |
| kim0 | waltc: the losetup and tgtadm would worry me | 16:59 |
| kim0 | the rest not really .. Java apps are chatty :) | 16:59 |
| kim0 | Are you actually able to start isntances and connect to ebs volumes | 17:00 |
| waltc | as I had said earlier...I blew away the logs (they do get massive) and did a fresh restart | 17:00 |
| === dendrobates is now known as dendro-afk | ||
| waltc | no....its a function of when the author choses to issue the log.info(...). I actually don't mind info (I use them all the time) but they can be overused thus lowering the 'signal to noise' ratio | 17:01 |
| === dendro-afk is now known as dendrobates | ||
| waltc | what I am trying to ascertain is for the real issues I have, are their roots during startup | 17:02 |
| waltc | you know...separate cause from effect | 17:03 |
| kim0 | yeah .. | 17:04 |
| waltc | I don't have iscsi so the admtgt I don't think should even be an issue. Well, the other thing too, are these benign, like the iscsi one | 17:04 |
| waltc | the loop device....that worries me too | 17:04 |
| waltc | at first, I tried to do things outside the 'script' like not have the 'everything but node controller' on my desktop because it saved buying a dedicated machine and it seemed like it 'should' work | 17:06 |
| waltc | largely it did except for disappearing public ipaddrs | 17:06 |
| waltc | with 10.10 I now have 3 machines, a dedicated Ubuntu server for 'everything but' and a dedicated ubuntu server for the nc | 17:07 |
| waltc | I did a free install...out of the box, manually config'd eth1 | 17:08 |
| waltc | works GREAT for store images of ubuntu. does not work for anything else | 17:08 |
| waltc | frankly, what I am trying to do kim0 is get buy-in at work to put up UEC there for my dev team, but I don't want egg on my face as soon as someone wants to start a centos image and can't | 17:10 |
| kim0 | waltc: have you tried publishing a centos image | 17:11 |
| kim0 | that should work AFAIK | 17:11 |
| waltc | I really want to get this to be predictable and reliable...and, as I said to you earlier, become really well versed in this...local 'expert' and all | 17:11 |
| waltc | I did try an image published by Eucalyptus | 17:12 |
| waltc | one commonality is the lack of a ramdisk image...but that should not necessarily cause a failure | 17:12 |
| waltc | the image starts, just never transitions to running | 17:12 |
| waltc | since it is not running I can't get to a log | 17:13 |
| waltc | with maverick and naddy images do I need to pull them apart and publish the individual parts separately? They run great under kvm natively | 17:14 |
| kim0 | waltc: can you try with this image http://eucalyptussoftware.com/downloads/eucalyptus-images/euca-centos-5.3-x8 | 17:14 |
| kim0 | 6_64.tar.gz | 17:15 |
| kim0 | broken link :) | 17:15 |
| waltc | just publish tarball? | 17:15 |
| kim0 | yeah mostly | 17:15 |
| kim0 | waltc: Also check out this message http://www.mail-archive.com/ubuntu-cloud@lists.ubuntu.com/msg00367.html | 17:17 |
| kim0 | where smoser mentions an updated patch to the uec-publish-tarball utility | 17:17 |
| kim0 | which may help | 17:17 |
| waltc | I will check it out...What I did for centos is pull it apart, remove the xen subdir moved the kvm subdir up a level and re-tared it | 17:22 |
| waltc | didn't go to run state | 17:22 |
| waltc | in other words, in images that work, there is no kvm subdir, the kernel manifest and ramdisk image are all in the single subdir | 17:23 |
| waltc | while centos is on my list...I don't want to shotgun the issues...so I am content, for now, resolving why maverick and naddy uec images from uec-ubuntu don't 'just work' | 17:25 |
| waltc | there was that one forum problem someone entered where he said he got frustrated and wiped the 'everything but nc' server and started from scratch to reinstall (plus added a new user) and now his works. | 17:26 |
| waltc | I don't find 'reformat and reinstall' a satisfying solution | 17:27 |
| waltc | there is code in walrus land that differentiates administrator from not administrator but that is just for permissions | 17:29 |
| waltc | is the focus on having these images work on ec2 and 'the hope is' they also will work on uec? | 17:30 |
| kim0 | waltc: hey, are you "wcorey" from forums | 17:32 |
| waltc | kim0, do you have a maverick based uec cloud where all images store or otherwise just work?....yes I am | 17:32 |
| kim0 | waltc: nice meeting you man :) | 17:32 |
| waltc | you too! | 17:32 |
| kim0 | I don't currently have UEC running with me | 17:33 |
| waltc | currently is this second or post 10.10? | 17:33 |
| kim0 | but yeah image store images, should definitely just work | 17:33 |
| kim0 | this second | 17:33 |
| waltc | my question is not 'should' they work it is 'do' they work | 17:33 |
| kim0 | when I played with 10.10, they did boot | 17:34 |
| waltc | the 'store' images do run perfectly...the maverick and naddy I get off the ec2/uec site start just fine but never go to running | 17:34 |
| kim0 | hmm | 17:35 |
| kim0 | well, with no particular error message | 17:35 |
| kim0 | it's gonna be a bit hard | 17:35 |
| kim0 | did you try nc logs ? | 17:35 |
| waltc | did you ever look at forum, t=1588826? | 17:36 |
| waltc | yes I did | 17:38 |
| waltc | to re-state what I believe you said is you have (somewhere) a UEC cloud, based on 10.10 and you downloaded the image from the ec2/uec site of maverick and naddy and published them uec-pub-tarball and they run just fine. Is that accurate? | 17:39 |
| kim0 | yeah, just not natty, it wasn't there yet | 17:40 |
| waltc | now I feel ill...that means the problem is here | 17:40 |
| kim0 | waltc: it might have needed some manual step though, can't quite remember | 17:41 |
| waltc | oh...do you run them as you or admin? | 17:41 |
| kim0 | waltc: man, I understand you want to help UEC users "debug" such problems | 17:41 |
| kim0 | waltc: as opposed to the current format/reinstall appraoch | 17:41 |
| kim0 | is that correct | 17:41 |
| waltc | btw, I enjoy your 'blogs' like publishing what scott did with the running the cloud-init under kvm directly | 17:42 |
| kim0 | waltc: What about we (you and I) start like some project to freshly install UEC, documenting the exact setup | 17:43 |
| kim0 | waltc: with every command ever needed ...etc | 17:43 |
| kim0 | if we stumble across an issue, we can hopefully get some help from the devs over here | 17:43 |
| kim0 | the advantage being .. | 17:43 |
| kim0 | instead of being an individual support case | 17:44 |
| kim0 | it would become .. like an open UEC manual for the whole world | 17:44 |
| waltc | in that case it wasn't so much trying to help him, he already got a solution. What I was trying to do is either from him or someone like yourself, explain which part of what he did actually fixed the problem. I logged in as admin, and just downloaded the credentials there so maybe I should not use admin just add myself as joe user and run images that way | 17:44 |
| kim0 | Everything you wanted to know about UEC but were afraid to ask .. comes to mind :) | 17:44 |
| kim0 | waltc: the thing is .. he doesn't know, nor do I .. and if you ask the devs .. so many things could go wrong | 17:45 |
| kim0 | there is no simple answer | 17:45 |
| kim0 | which is why I am suggesting we start documenting a "controlled" envrionment | 17:45 |
| kim0 | I would personally love to work on this with you | 17:46 |
| waltc | kim0....yes, yes, and yes | 17:46 |
| kim0 | just let me know if you can commit to doing it | 17:46 |
| waltc | I've been in software, as a developer for decades | 17:46 |
| waltc | but part of that is support | 17:46 |
| waltc | what, as a person paid to do this stuff, really annoys me is people who open a problem, find some solution and just simply announce, "OK, thanks, it's fixed now" without ever saying what they did | 17:47 |
| kim0 | hehe yeah .. some tend to do that :) | 17:48 |
| waltc | the next worst thing is people that say, I did this, this, this, this, and this....and now it works | 17:48 |
| waltc | that, as he described it, is the "hail mary" solution... | 17:48 |
| waltc | ALOT tend to do that | 17:49 |
| waltc | I like crisp and clean answers..THIS (and only this) is the cause and the solution is do this (and only this) | 17:49 |
| waltc | link cause and effect and remedy together | 17:50 |
| waltc | Oh...I have become an expert at re-installing, from scratch, the ubuntu server (as cluster) and the ubuntu server as node | 17:51 |
| kim0 | waltc: I am in total agreement | 17:51 |
| kim0 | about the "this and only this" statement | 17:51 |
| kim0 | thus I am suggesting we start working on a controlled env | 17:52 |
| kim0 | documenting everythgin aloong the way | 17:52 |
| * kim0 is on crappy Internet for now .. slow answer | 17:52 | |
| waltc | the other thing I hate is people who open an issue and it gets kidnapped by others who say I have the same problem except... After a couple of those you don't even know what the problem environment even is any more | 17:53 |
| waltc | I realize to most, if not all, people who work on the forums and maybe even launchpad, theirs is a 'labor of love' in other words they are there because they are interested in being there, not they have to be. So when a problem comes in and it is, or becomes by others additions, too confusing they move on as it is now too confiusing | 17:55 |
| waltc | let's start here...something I wanted to verify with you....10.4 is LTS, 10.4 UEC is 1.6, 10.4 won't run Eucalptus 2 or UEC 2.0 as it needs the newer kernel. What you get with UEC 2 is the cloud init stuff which means you don't have to specialty create new images for everything but do stem cell image that gets specialized at boot time, not create time...correct? | 17:58 |
| waltc | or, asked differently, why would one want to use UEC 2.x? | 17:59 |
| kim0 | waltc: 10.04 is LTS and runs 1.6 yes .. can it run 2.0, maybe | 18:00 |
| kim0 | cloud-init is not directly related to EUC version AFAIK .. it's an image thing | 18:00 |
| waltc | I seem to recall a forum question and answer saying you could not install UEC 2 as Eucalyptus 2.x required 2.6.35 | 18:01 |
| TeTeT | waltc: there's linux-image-generic-pae-lts-backport-maverick for lucid in the repo | 18:01 |
| TeTeT | I guess an backport of euca 2 for lts would be an interesting thing to do | 18:02 |
| waltc | but cloud init IS a Ubuntu thing...until very recently you couldn't use it on non ubuntu images. Amazon recently released a centos image that does support cloud init...right? | 18:02 |
| kim0 | TeTeT: any idea if euca2 does need a newer kernel | 18:02 |
| TeTeT | kim0: actually not | 18:02 |
| kim0 | waltc: ^^ thought so | 18:02 |
| kim0 | waltc: yeah amazon released their own "Linux distro" basically | 18:03 |
| waltc | weird huh? | 18:03 |
| kim0 | waltc: and they did choose to include cloud-init | 18:03 |
| waltc | that pretty much locks Amazon into being a distro as unless RH includes cloud-init it won't be part of stock centos | 18:05 |
| waltc | correct? | 18:05 |
| kim0 | Amazon's Linux, IS a separate distro | 18:05 |
| kim0 | and yes .. centos will only get it when RH acccepts it (if ever) | 18:05 |
| kim0 | unless someone packages it for centos directly | 18:06 |
| waltc | you mean all Amazon created AMI's or just that centos with cloud-init? | 18:06 |
| \sh | moins | 18:07 |
| waltc | why wouldn't they (RH)...just like why wouldn't Ubuntu/Canonical include Spice? | 18:07 |
| \sh | hmmm...when I want to upgrade an ec2 lucid instance to maverick, and I did a dist-upgrade from lucid to maverick...what aki id do I need to set to boot the correct kernel (which should be linux-image-2.6.35-24-virtual) | 18:07 |
| kim0 | waltc: It's not like Ubuntu is opposed to spice .. It's just barely mature right now .. I'm sure Ubuntu will get it (personal opinion) | 18:08 |
| kim0 | waltc: it's a matter of opinion .. if RH wants it, it can get it .. it's all open source | 18:08 |
| kim0 | \sh: did you try "ec2-describe-images" | 18:11 |
| kim0 | \sh: it should tell you the associated aki/ari | 18:11 |
| \sh | kim0: yes, I followed smosers howto http://ubuntu-smoser.blogspot.com/2010/04/upgrading-ebs-instance.html | 18:11 |
| \sh | but I don't find the kernel version at all... | 18:11 |
| waltc | We recently went through an exercise at work to put much of our infrastructure in EC2. building specialized images are a pain in the butt, slow, fragile, etc..I MUCH prefer the stem-cell provision on boot approach | 18:11 |
| \sh | during dist-upgrade it installed the 2.6.35-24-virtual flavor | 18:11 |
| kim0 | waltc: Yep! :) | 18:12 |
| waltc | between delta-cloud and RHEV RH has staked it's 'claim' on the cloud and virtualization space....it stands to reason they (and SUSE) would grab cloud-init | 18:13 |
| waltc | I would think the only incentive not to is to 'one-off' someone else's distro | 18:14 |
| kim0 | \sh: smoser 's blog post seems to nail it | 18:14 |
| \sh | kim0: he was upgrading to lucid...and I'm doing dist-upgrade from lucid to maverick, so I adjusted his shell/ec2-api-tools magic to the kernel version I got during dist-upgrade | 18:15 |
| \sh | kim0: but this version I can't find the version in the ec2-describe-images --all list | 18:15 |
| \sh | kim0: the only kernel which matches on the list is IMAGEaki-2316fd4a099720109477/ububtu-test-kernel-2.6.35-12-virtual-x86_64-20100730-test-10 | 18:18 |
| kim0 | \sh: The 64 bit aki seems to be aki-427d952b, 32bit aki-407d9529 | 18:18 |
| smoser | shang, what kernel do you have installed | 18:19 |
| smoser | and what release (maverick?) are you running ? | 18:20 |
| kim0 | smoser: he's upgrading lucid to maverick | 18:20 |
| \sh | smoser: I installed the t1.micro ebs image of lucid...and dist-upgraded to maverick now | 18:20 |
| \sh | smoser: the running instance kernel is linux-image-2.6.32-311-ec2 and the dist-upgraded kernel now is linux-image-2.6.35-24-virtual | 18:21 |
| smoser | ah. then, yeah, you want the pv-grub kernel. make sure that you install grub-legacy-ec2 | 18:21 |
| \sh | (hence no linux-image...-ec2 package for maverick) | 18:21 |
| \sh | the grub-legacy-ec2 is already installed | 18:22 |
| \sh | smoser: and what is the pv-grub kernel? | 18:22 |
| smoser | yeah, then you have a /boot//grub/menu.lst that lists | 18:22 |
| smoser | linux-image-2.6.35-24-virtual | 18:22 |
| smoser | kim0 listed the pv-grub kernel above | 18:22 |
| \sh | smoser: ah ok... | 18:23 |
| kim0 | \sh: this actually chainloads the kernel inside the image .. which is cool | 18:23 |
| kim0 | smoser: do you actually memorize the akis ;) | 18:23 |
| \sh | so I just update via modify-instance the kernel id to the aki mentioned. | 18:23 |
| smoser | no, i look them up. but i do have cached describe-images of all regions | 18:24 |
| smoser | yes, \sh | 18:24 |
| kim0 | smoser: um, once on pv-grub .. upgrading maverick->natty .. should just involve dist-upgrade, like a bare metal box right ?! | 18:24 |
| kirkland | kim0: when does that video go up? | 18:25 |
| smoser | thats the goal, kim0 yes | 18:25 |
| smoser | kirkland is looking for his 15 minutes of youtube fame | 18:26 |
| kim0 | kirkland: :) Jono is holding it hostage .. should definitely go up today ... | 18:26 |
| * kim0 will bug Jono as soon as I see him | 18:26 | |
| \sh | hmmm | 18:27 |
| \sh | I think I destroyed the instance, grmpf...so now I directly start a maverick instance | 18:28 |
| kim0 | :) | 18:28 |
| kim0 | \sh: I hope you didn't loose any data on this :) | 18:29 |
| smoser | \sh unless you terminatede the instance, you didn't loose its root volume | 18:30 |
| smoser | (as it is ebs) | 18:30 |
| \sh | kim0: no...it's just this t1.micro free amazon offer | 18:30 |
| kim0 | \sh: stopping is not terminating | 18:31 |
| smoser | kim0, you or i could walk the lucid -> maverick upgrade path in a blog post | 18:31 |
| smoser | that might actually be useful | 18:31 |
| \sh | kim0: terminated :) | 18:31 |
| kim0 | smoser: Sounds good :) | 18:32 |
| * kim0 likes | 18:32 | |
| kim0 | I just wish people would stop using ebs images .. it's a cloud c'mon :) | 18:33 |
| * kim0 notes that blog entry | 18:34 | |
| kim0 | smoser: I suppose Amazon could update their PVgrub kernel .. how can we always find out the latet ID ? | 18:34 |
| smoser | kim0, no way. | 18:40 |
| * kim0 #&$^%$^$ @amazon | 18:40 | |
| smoser | other than look at the kernel associated with our latest image of that release/arch | 18:40 |
| kim0 | Yeah ok | 18:40 |
| crazed | how do you enable root logins in an ec2 ami? | 21:14 |
| crazed | i changed /etc/cloud/cloud.cfg disable_root: false | 21:15 |
| crazed | but still seems to put in the "please login as ubuntu" | 21:15 |
| waltc | kim0: I actually have fully documented the installation process and post installation I did this last time around. It appears managed-novlan requires a second eth on the CC (in a default UEC install) yet it doesn't configure it. In fact, if it doesn't exist, it doesn't care. | 21:24 |
| waltc | that's what I was missing in the 9.04->10.04 time frame. BTW, my old issue of disappearing public ip addresses went away perhaps because of 10.10 or perhaps because of eth1 | 21:26 |
| kim0 | kirkland: Your vid is out and loud woohoo http://www.youtube.com/ubuntudevelopers#p/a/u/0/CWNjzwci5e8 | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
