Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-03-19

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:12 aglarendil joined #fuel
00:12 miroslav_ joined #fuel
00:13 mattymo joined #fuel
00:18 aglarendil|pc joined #fuel
01:19 rmoe joined #fuel
01:49 crandquist joined #fuel
02:33 fandi joined #fuel
02:49 xarses joined #fuel
03:55 vkozhukalov_ joined #fuel
04:29 mihgen joined #fuel
04:42 IlyaE joined #fuel
05:28 dburmistrov joined #fuel
05:32 dburmistrov joined #fuel
05:58 saju_m joined #fuel
06:05 IlyaE joined #fuel
06:49 Ch00k joined #fuel
07:08 mihgen joined #fuel
07:27 Ch00k hi guys! I made a deployment with fuel and when running health check for Heat it says " Is not implemented for this deployment mode"
07:28 Ch00k what exactly does this mean?
07:52 e0ne joined #fuel
08:22 rvyalov joined #fuel
08:37 evg joined #fuel
08:47 warpig joined #fuel
08:48 Ch00k joined #fuel
08:56 topochan joined #fuel
09:03 e0ne joined #fuel
09:04 mihgen joined #fuel
09:14 alex_didenko joined #fuel
09:22 e0ne joined #fuel
09:46 e0ne_ joined #fuel
09:50 warpig morning guys...
09:50 warpig MiroslavAnashkin suggested I run the updater to 4.1.1 to resolve HP disk controller issues with a Ubuntu deployment
09:50 warpig I did this, unfortunately, the installation still cannot see the disks...  :o(
09:51 warpig there is no /dev/hd* /dev/sd* /dev/cciss*
09:52 warpig lspci shows "03:00.0 RAID bus controller: Hewlett-Packard Company Smart Array Gen8 Controllers (rev 01)
09:52 warpig and I've confirmed there is a logical drive configured.
09:53 Ch00k joined #fuel
09:55 warpig the bootstrap kernel correctly detects everything (but I believe that's based on CentOS)
10:17 e0ne joined #fuel
10:21 Arminder can someone have a look at -> https://bugs.launchpad.net/fuel/+bug/1275754
10:33 vkozhukalov_ joined #fuel
10:37 Ch00k_ joined #fuel
10:37 acca1 joined #fuel
10:38 acca1 left #fuel
10:45 Ch00k joined #fuel
10:49 mirrorbox joined #fuel
10:50 mirrorbox hi
10:50 mirrorbox I follewed the instruction here: http://docs.mirantis.com/fuel-dev/develop/env.html?highlight=fuel%20web
10:50 mirrorbox and tests fail for me
10:52 mirrorbox ah, nvm, missed createdb step
10:52 anotchenko joined #fuel
10:56 rvyalov joined #fuel
11:03 anotchenko joined #fuel
11:24 anotchenko joined #fuel
11:52 e0ne joined #fuel
12:07 e0ne joined #fuel
12:10 e0ne_ joined #fuel
12:21 anotchenko joined #fuel
12:28 e0ne joined #fuel
12:30 evg left #fuel
12:30 evg joined #fuel
12:34 crandquist joined #fuel
12:54 TVR_ but.. but... that's more typing now...
12:54 TVR_ it's OK MiroslavAnashkin ... your expertise is worth the extra keystrokes...
12:56 ykotko joined #fuel
13:00 justif joined #fuel
13:01 evgeniyl joined #fuel
13:30 evgeniyl left #fuel
13:30 evgeniyl joined #fuel
13:45 Ch00k joined #fuel
14:04 anotchenko joined #fuel
14:08 vkozhukalov_ joined #fuel
14:34 miroslav_ joined #fuel
14:40 jobewan joined #fuel
14:47 MiroslavAnashkin warpig: Please remove the old environment, existing before the patch, create new one, add nodes to the new env,
14:47 warpig MiroslavAnashkin: yep, I did that - still no love
14:48 warpig I rebooted the Fuel node too
14:48 MiroslavAnashkin warpig: Then, check if your disks are visible in Fuell UI and have correct capacity. If no - create diagnostic snapshot. If yes - try to deploy first and create diagnostic snapshot in case of deployment failure.
14:49 warpig do you have an MD5 of the patched initrd.gz so I can verify I'm installing the correct image?
14:50 warpig Within Fuel, all disks are reported correctly.  My controller node has 146GB in RAID1 and my two compute nodes have 300GB in RAID0
14:50 warpig I've gone with a CentOS installation for now - that's working fine
15:01 IlyaE joined #fuel
15:05 anotchenko joined #fuel
15:10 miroslav_ warpig: One issue with all these patches - no one patch adds updated packages to local repositories, existing on master node
15:11 miroslav_ So, as soon as you successfully deployed the env, all deployed nodes use old agent to report status to master node.
15:32 warpig miroslav_: but the disk discovery would come from the boot kernel at that point, not the installed package kernel
15:33 miroslav_ Yes, but Mirantis Openstack uses separate agent to report to master node.
15:34 miroslav_ So, old agent may report there is no disks while actually everything is OK
15:38 warpig miroslav_: I'm not sure I follow completely...
15:38 warpig when my nodes boot into the bootstrap kernel, they report everything correctly to Fuel - no problems there.
15:39 warpig When I create an environment based on Ubuntu and deploy, that is when they fail.
15:39 warpig If I create a different environment based on CentOS, they deploy without a problem.
15:40 miroslav_ warpig: OK. Then please create and share a new diagnostic snapshot after Ubuntu installation failure.
15:41 warpig I'm not sure how much use that would provide, as all the hosts are stuck in a loop when trying to partition a non-existent disk...
15:42 warpig If you can get any info about the hardware, I'm happy to provide one.
15:42 anotchenko joined #fuel
15:50 ggreenwalt1 Anyone had a PXE booted system sit on "child_rip+0x0/0x20" before?
15:50 ggreenwalt1 I'm wanting to do a test depoyment but I can't get an installation on an unallocated node to progress pas that.
15:53 e0ne joined #fuel
15:54 warpig ggreenwalt1: what part of PXE does it stop?  Does it manage to load the kernel?
15:54 anotchenko joined #fuel
15:55 miroslav_ warpig: Yes, diagnostic snapshot provides info about hardware, GRUB settings etc
15:58 miroslav_ ggreenwalt1: Do you mean you unable to add node to env since it does not get boot to PXE bootstrap image?
15:58 ggreenwalt1 It boots to the grub menu.
15:59 ggreenwalt1 And then when I hcoose "local" (or ignore it for 5 seconds) it begins the bootstrap startup.
15:59 ggreenwalt1 And then during the boot it hangs on child_rip.
16:00 miroslav_ ggreenwalt1: Do not choose anything from boot menu. Simply wait 5 seconds
16:00 ggreenwalt1 Vm Power -> PXE -> Automatic Boot in 1 second...    \\ Loading /images/bootstrap/linux ......
16:01 miroslav_ This menu is enabled for those who want to use custom bootstrap image.
16:01 ggreenwalt1 then a bunch of stuff loads. ala    [<ffffff8111c918>] ? generic_file_aio_write +0x88/0x100
16:01 angdraug joined #fuel
16:01 ggreenwalt1 and finally it settles on "[<ffffffff8100c0c0>] ? child_rip+0x0/0x20
16:02 ggreenwalt1 the CPU then sits at 8% indefinitely.
16:03 miroslav_ ggreenwalt1: Oh, then tell us a bit about your hardware
16:04 ggreenwalt1 Well in this instance it's Hyper-V running on a 3930k i7.   32GB of RAM.. everything else I imagine is virtualized so the Hyper-V hypervisor is the most significant "hardware".
16:06 miroslav_ Hmm. Never tested on Hyper-V. Fuel supports VirtuslBox and KVM out of the box and VMWare after some manual settings
16:06 miroslav_ VirtualBox
16:08 warpig ggreenwalt1: I haven't used Hyper-V, but I assume it wouldn't install without VT enabled in the BIOS?  Something to check?
16:09 ggreenwalt1 Yeah, ultimately it needs to peacefully co-exist on a Hyper-V 2012 server.   I tried enabling Hyper-V's failsafe CPU settings but no luck.   I'll open a ticket with Mirantis but I thought it might be a common problem with an easy workaround.  The only mention on the Googles I found of this problem were people installing CentOS and needing a new kernel update.  But I suspect the kernel for Fuel is pretty finely tuned and would break all mann
16:09 miroslav_ Fuel uses CentOS kernel for master node(
16:11 ggreenwalt1 Yeah which is what's weird.  I can isntall the master node in Hyper-V no problem.
16:11 miroslav_ We are working on changing kernel to recent Fedora one for bootstrap and masternode. There are drivers, which are not exist for 2.6 kernel
16:15 Llew joined #fuel
16:15 Llew Hello.
16:21 anotchenko joined #fuel
16:31 miroslav_ Hello.
16:33 anotchenko joined #fuel
16:34 ggreenwalt1 Well, back to work.  Good luck with your systems.  Cheers.
16:34 ggreenwalt1 left #fuel
16:42 miroslav_ joined #fuel
17:04 topochan joined #fuel
17:08 dburmistrov joined #fuel
17:17 evgeniyl` joined #fuel
17:21 ggreenwalt joined #fuel
17:21 ggreenwalt Just a heads up I solved my Kernel Panic problem on the PXE boot by disabling dynamic memory in Hyper-V and increasing the startup RAM value.
17:22 ggreenwalt So if anyone ever wants to run on Hyper-V and has that problem. it's an easy fix.
17:23 rmoe joined #fuel
17:24 BillTheKat joined #fuel
17:25 Gil_McGrath I added in ssh public key via the Fuel GUI - did not show up in /root/.ssh/authorized_keys file after provisioning. Are there any known issues relative to this?
17:26 Ch00k joined #fuel
17:28 anotchenko joined #fuel
17:32 Gil_McGrath joined #fuel
17:39 MiroslavAnashkin Hmm, no. it is new issue
17:43 ggreenwalt left #fuel
17:45 jseutter joined #fuel
17:51 anotchenko joined #fuel
17:55 IlyaE joined #fuel
17:56 Gil_McGrath I will put in a support ticket
17:59 MiroslavAnashkin To file a bug here https://bugs.launchpad.net/fuel works better
18:09 Gil_McGrath done
18:10 Ch00k joined #fuel
18:10 Gil_McGrath now there are both a bug report and support ticket
18:11 MiroslavAnashkin Thank you!
18:12 Gil_McGrath by the way - where are the nova-api, glance-api etc log files - I do not see nova-api logfile in /var/log/nova
18:13 MiroslavAnashkin All logs are on master node in /var/log/remote
18:14 Gil_McGrath really?  The Fuel node is needed..... (i.e. critical in operations)
18:14 Gil_McGrath thanks for the info
19:04 e0ne joined #fuel
19:13 designated joined #fuel
19:22 saju_m joined #fuel
19:23 vkozhukalov_ joined #fuel
19:27 designated is there currently an option in fuel 4.1 to load host OS across multiple drives using mdadm to account for a lack of hardware raid controllers?  Or are you pretty much limited to a single physical drive for host OS?
19:41 fandi joined #fuel
19:51 mutex designated: there is nothing in the fuel UI for 4.0, not sure about 4.1
19:52 mutex but if you use the fuel JSON API you can craft all those kinds of rules for the installer
19:52 mutex just no gui for it
19:52 mutex its in the customizing fuel documentation
19:52 mutex I have not done it, but it looks easy
20:07 crandquist joined #fuel
20:08 saju_m joined #fuel
20:11 designated joined #fuel
20:54 e0ne joined #fuel
21:00 rvyalov joined #fuel
21:06 albionandrew joined #fuel
21:08 albionandrew Hi, we have a cluster that has a compute node we need to replace the mother board on. Is it better to delete that node? Reprovision or better to change the mac addresses manually?
21:20 IlyaE joined #fuel
21:39 albionandrew joined #fuel
21:39 albionandrew How long should I have to wait for a node to be deleted/removed from the cluster? The node in question has said pending deletion for 15 mins.
21:45 designated albionandrew, I've had to wait for a while before, eventually it succeeded.
21:46 albionandrew an hour?
21:47 designated albionandrew, i don't know, I got tired of waiting so I went to eat dinner, came back and it was finished.
21:47 albionandrew ok Thanks. I'll see what happens
21:50 rvyalov joined #fuel
22:04 e0ne joined #fuel
22:51 xarses joined #fuel
23:26 rmoe joined #fuel
23:58 IlyaE joined #fuel

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary