Camelia, the Perl 6 bug

IRC log for #fuel, 2013-10-29

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

All times shown according to UTC.

Time Nick Message
00:45 rmoe joined #fuel
01:08 xarses joined #fuel
01:11 tonyha joined #fuel
01:15 tonyha Hi on fuel 3.2 is there away to seperate the fuel web url ip from the pxe lan using 2 nic card ?
01:23 xarses yes, if you have an additional network it can be configured and can be used separately
01:25 xarses but neither should intersect with the networks used in each cluster network settings
01:25 tonyha I tried using 2 nic card but when I try to connect it give ssl error on the browser.  This is under vbox
01:27 tonyha is /usr/local/sbin/bootstrap_admin_node.sh the correct tools to change the ip and interface ?
01:29 xarses the fuel server is on port 8000
01:29 xarses ssl cert is self signed, so that kind of error is expected
01:29 xarses apache on 80, and 443 is used by cobbler
01:30 tonyha Going to 80 it give apache site going to 443 it will ask to accept certificate.  Going to 8000 give ssl error
01:31 xarses check https:, should be http://<ip>:8000/
01:32 tonyha oh ok that is the error.  Thank you
01:32 xarses you welcome
01:33 tonyha One more thing is there a way to use the latest 2013.1.4 build of openstack from fuel or need to rebuild the iso for that ?
01:34 xarses deb or yum?
01:35 tonyha yum
01:35 xarses its easy enough to add them to the repos in /var/http/nailgun/
01:36 xarses and and do create-repo -g repodata/comps.xml --update
01:36 xarses from the repo folder
01:37 tonyha ok will try that. Thank you
01:37 xarses the deb packages are much harder to update in fuel node since it's centos and dosn't have deb repo tools
01:37 xarses in that case it's easier to rebuild iso usually
01:38 tonyha Ini production are there pro/cons using ubuntu vs centos ?
01:43 xarses Mostly it's about which your organization might prefer from a support perspective and if you're already a so-n-so shop
01:43 xarses there is a trade off between RHEL distros tending to be more stable and Debian distros having more features
01:44 rmoe joined #fuel
01:45 tonyha ok.  Looking at /var/www/nailgun/centos/fuelweb/x86_64/Packages  does the rpm with mira0 or mira1 identical with normal build rpm from fedora ?
01:46 tonyha so to use the latest build just replace all the 2013.1.3 with 2013.1.4 build from fedora repo ?
01:50 xarses thats about it
01:50 xarses you do need to rebuild the repodata and not forget to add the comps.xml back
01:50 xarses or the package groups wont work on the kickstart
01:51 xarses as to the difference, if it includes miraX its because we rebuilt the package for some reason. Usually to add special patches
01:53 xarses likely because we needed to add something that wasn't accepted upstream yet
01:53 tonyha almost all openstack rpm have the miraX :)  I hope changing to 2013.1.4 will not break fuel deployment.
01:54 xarses it could
01:55 xarses it probably wont, but it might have some runtime issue that we added a patch for
01:55 tonyha on the fuel-dev environment does it show what are being patch for the miraX rpm ?
01:58 tonyha or it just download from mirantis repo ?
01:59 xarses I'm not sure how the change log is kept
01:59 xarses they are uploaded to download.mirantis.com from our package team
02:00 xarses beyond that i'd probably have to inspect each package
02:04 tonyha by adding the 2013.1.4 release and updating the comps.xml file does it make it available on the Releases windows on fuel.  So we can choose between 2013.1.3 and 2013.1.4 ?
02:07 xarses no
02:07 xarses it makes it default for centos install
02:08 tonyha ok.  Will try.  Thank you for your help.
03:55 tonyha joined #fuel
03:55 Arminder joined #fuel
03:56 Arminder [28-Oct-13 10:33] <ArminderS> anyone taking care of my ticket -> http://support.mirantis.com/requests/964
03:56 Arminder [28-Oct-13 10:34] <ArminderS> creating volume from image, with ceph for images & volumes, isn't working
03:57 ArminderS joined #fuel
03:59 ArminderS- joined #fuel
04:06 AndreyGrebenniko joined #fuel
04:07 AndreyGrebenniko Hi all! Nice to get a channed divided
04:27 xarses joined #fuel
04:27 xarses_ joined #fuel
05:41 tonyha was watching http://www.youtube.com/watch?v=Kc6SzONqOYk  on OpenStack Heat and autoscaling demo.
05:41 tonyha She mention about auto scaling template in github.  Anybody know what the url is ?
05:45 ArminderS https://github.com/openstack/heat-templates
05:46 ArminderS https://github.com/openstack/heat-templates/co​mmit/85f5a12ab418d564863bb42637c350622f3f2cf7
06:06 MiroslavAnashkin joined #fuel
06:06 mattymo joined #fuel
06:06 meow-nofer joined #fuel
06:06 xdeller joined #fuel
06:07 evgeniyl joined #fuel
06:28 ArminderS joined #fuel
06:29 MiroslavAnashkin joined #fuel
06:45 ArminderS i wish to confirm that all VM traffic goes on br-int when we are using neutron with gre implementation?
06:45 ArminderS can someone confirm that for me?
06:47 mihgen joined #fuel
06:47 ArminderS does it contain traffic for vm only or other openstack traffic as well like api & services traffic like for volumes/images/objects/etc.?
06:47 ArminderS if not, where does the rest of traffic could be catched from?
06:54 mihgen Arminder: other openstack traffic goes over "management" network in Fuel, so it's always separated from VMs
06:54 mihgen http://docs.mirantis.com/fuel/fuel-3.2/refere​nce-architecture.html#neutron-vs-nova-network this pics may help to understand it better
06:54 mihgen there are stylistic issues though..
06:55 ArminderS these 2 were posted befroe you joined, so pasting again
06:55 ArminderS [29-Oct-13 12:15] <ArminderS> i wish to confirm that all VM traffic goes on br-int when we are using neutron with gre implementation?
06:55 ArminderS [29-Oct-13 12:15] <ArminderS> can someone confirm that for me?
06:56 ArminderS so br-int is basically the private network, right?
06:58 mihgen ArminderS: I can't say for sure, sorry :) We need to wait neutron experts to show up or you can do ovs-vsctl show to see what goes where..
07:00 ArminderS right
07:07 ArminderS i understand br-mgmt is for mgmt traffic, br-storage for storage traffic
07:07 ArminderS just need to confirm what goes in/out of br-ex, br-int, br-prv & br-tun
07:17 mihgen ArminderS: if you stay up for ~2h, I'll get someone to respond on your question… not everyone is here yet
07:18 ArminderS sure, not an issue
07:18 ArminderS i'm her for another 4 hours or so
07:18 ArminderS here*
07:20 e0ne_ joined #fuel
07:38 ogelbukh joined #fuel
07:44 e0ne joined #fuel
07:51 e0ne joined #fuel
08:17 amartellone joined #fuel
08:42 ArminderS anyone from mirantis checking my support request -> http://support.mirantis.com/requests/964
08:42 ArminderS haven't heard back since 26/oct
08:52 bogdando joined #fuel
08:57 jkirnosova joined #fuel
09:19 archon joined #fuel
09:22 mihgen joined #fuel
09:37 teran joined #fuel
09:39 teran joined #fuel
09:40 teran joined #fuel
09:42 figless joined #fuel
09:42 teran joined #fuel
09:43 sanek joined #fuel
10:05 teran joined #fuel
10:20 e0ne_ joined #fuel
10:26 Bomfunk joined #fuel
10:26 sanek joined #fuel
10:26 Bomfunk Hello mattymo! /join #fuel-dev
10:26 Bomfunk uhm. sorry for that.
10:26 Bomfunk mattymo: Mike Scherbakov said that you probably can help us with the following question:
10:27 Bomfunk One of our customers wants to install Red Hat using Fuel. He has the subscription.
10:27 Bomfunk But he says that his fuel-pm doesn't have an internet access.
10:27 Bomfunk So he's looking for a way to configure http-proxy, for example.
10:28 Bomfunk So, the question is: is that possible with Fuel? If yes, then how? Thanks.
10:42 amartellone Hi all, I'm running fuel-web in fake mode, but when I do manage.py loaddefault I've the follow error
10:42 mattymo Bomfunk, we don't have any support for Red Hat without internet access
10:42 amartellone http://paste.openstack.org/show/50114/
10:42 mattymo if you can access the internet via a simple http proxy, it might work but I can't promise
10:43 Bomfunk So... Should I tell the customer that it's impossible?
10:43 mattymo Bomfunk, you should modify /etc/puppet/modules/rpmcache/​tempaltes/build_rpm_cache.erb and add "--proxy=http://yourproxy.yourdomain.com:3128" to subscription-manager
10:43 mattymo or the customer could set up Red Hat Satellite Server
10:44 tonyha joined #fuel
10:44 mattymo this won't work because if we can't reach cdn.redhat.com directly from nailgun, it will mark the deployment as failed and won't even start
10:45 mattymo and Satellite will fail too because we need to download 8 packages from RHN before using Satellite as well
10:52 mihgen joined #fuel
10:54 mihgen amartellone: we have fuel-dev channel for such questions. Most likely you forgot to do ./manage.py syncdb  before loading daya
10:54 mihgen data*
10:58 lambda_ joined #fuel
10:59 mattymo #fuel-dev *
10:59 lambda_ left #fuel
11:01 vkozhukalov joined #fuel
11:02 amartellone sorry mihgen, now i switch on fuel-dev channel...but I've done how you say
11:06 mihgen meow-nofer: I need you in #fuel-dev
11:16 aabashkin joined #fuel
11:18 dburmistrov joined #fuel
11:18 dburmistrov #fuel-dev
11:20 b-zone joined #fuel
11:23 adanin joined #fuel
11:34 kcherchenko joined #fuel
12:02 vkramskikh joined #fuel
12:05 MiroslavAnashkin joined #fuel
12:07 MiroslavAnashkin joined #fuel
12:35 vkramskikh joined #fuel
12:37 vkramskikh joined #fuel
13:04 vkozhukalov joined #fuel
13:08 vkramskikh joined #fuel
13:25 r0mikiam joined #fuel
14:00 vk joined #fuel
14:36 r0mikiam joined #fuel
14:38 r0mikiam joined #fuel
15:03 teran joined #fuel
15:22 dpyzhov joined #fuel
15:24 dhblaz joined #fuel
15:26 dhblaz xarses:  I modified the kickstart on my fuel master so it installes the 3.10 kernel you asked me to install.  The machines that are BL280c G5 kernel panic on boot.  These machines are also unique in that they have p711m raid card.
15:26 dhblaz Sorry they are BL280c G6s
15:27 dhblaz Andrew is going to boot them this morning with serial console to capture the stack trace.
15:27 dhblaz When I look at the elrepo I don't see another 3.10 kernel just 3.11 kernels
15:35 xarses hmm, I'll speak with the engineer abut it when i get into the office
15:37 dhblaz Thanks
16:12 e0ne joined #fuel
16:19 angdraug joined #fuel
16:27 e0ne joined #fuel
16:27 IlyaE joined #fuel
16:36 alex_null joined #fuel
16:42 AndreyDanin joined #fuel
16:53 AndreyDanin_ joined #fuel
17:01 xarses joined #fuel
17:03 rmoe joined #fuel
17:06 dpyzhov Building iso with havana: http://jenkins-product.srt.mirantis​.net:8080/view/Product%20and%20Comm​unity/job/fuel_main_iso/291/console
17:07 * dpyzhov building iso with havana: http://jenkins-product.srt.mirantis​.net:8080/view/Product%20and%20Comm​unity/job/fuel_main_iso/291/console
17:08 vkozhukalov joined #fuel
17:27 e0ne joined #fuel
17:28 e0ne joined #fuel
17:37 archon joined #fuel
17:38 angdraug joined #fuel
18:00 dhblaz joined #fuel
18:01 dhblaz xarses: any word?
18:01 teran joined #fuel
18:23 xarses dhblaz: I spoke with the eng about this the 3.11 kernels he tested don't have the fix needed, they are only in 3.10 at this point (for centos)
18:25 xarses dhblaz: any kernel 3.10.x kernel from fedora's EPEL should have the patch in it. I'll have to prod around for some more information regarding finding the actual patch needed.
18:26 dhblaz When I get a good copy of the kernel panic I'll try to map that to a known bug and see where it has been resolved.
18:26 xarses it sound's like we might need to build a kernel for your driver set though if the one we had isn't working.
18:26 xarses do you know what driver it was around?
18:26 dhblaz On another note, interfaces that come up with the names like rename0
18:27 dhblaz continue to be a problem and thwart deployment
18:27 xarses did the biosdevname=0 param get lost?
18:27 dhblaz This is something that I imagine your team will want to address in their maintenance release
18:27 xarses yes, it is
18:28 xarses Also, the issue is not supposed to be in the 3.8 ubuntu kernel we use
18:28 dhblaz Andrew tried that patch yesterday and I double checked his work last night.  We don't end up with a booting OS.
18:29 dhblaz I haven't had time to look into it too much and I don't know much about the ubuntu install process to troubleshoot it.
18:30 dhblaz Our prefered distro is Scientific Linux (after the drama and resulting CentOS taking so long to get 6 out the door)
18:32 xarses your nic's where broadcom correct?
18:32 dhblaz Do you expect biosdevname=0 in the bootstrap kernel arguments?  Because cobbler doesn't appear to show any kernel arguments for bootstrap.
18:35 xarses in /var/lib/tftpboot/pxelinux.cfg/default
18:36 xarses the args line for bootstrap should have biosdevname=0
18:36 xarses the file is generated by cobbler sync
18:37 xarses so any manual changes wont stick after fuel attemps to deploy a node
18:37 dhblaz The motherboard nics are Intel
18:38 dhblaz and the PCI nics are Emulex
18:38 dhblaz igb
18:38 dhblaz be2net
18:38 xarses cobbler system dumpvars --name=default | grep bios
18:38 xarses kernel_options : ksdevice=bootif lang=  biosdevname=0 url=http://10.108.0.2:8000/api locale=en_US text kssendmac priority=critical
18:39 dhblaz [root@fuel1 ~]# cobbler system dumpvars --name=default | grep bios
18:39 dhblaz kernel_options : ksdevice=bootif lang=  biosdevname=0 url=http://10.62.12.10:8000/api locale=en_US text kssendmac priority=critical
18:52 dhblaz Confirmed that biosdevname=0 is in the boot arguments in grub.  I'm not sure they are then when we PXE chainload (which is what happens in production systems right now)
18:53 dhblaz Looks like the kernel panic is related to this error:
18:53 dhblaz "hpsa: detect controller lockup"
18:58 xarses maybe this patch https://lkml.org/lkml/2011/10/26/304
19:00 dhblaz I believe that patch introduces the error message we are seeing.
19:00 dhblaz dmesg reports" HP HPSA Driver (v 2.0.2-1)
19:08 xarses hmm that patch was in mainline 3.2+
19:12 IlyaE joined #fuel
19:31 e0ne joined #fuel
19:41 jkyle joined #fuel
19:41 jkyle where can I file bugs for fuel?
19:52 angdraug jkyle: https://launchpad.net/fuel
19:54 jkyle thanks
19:59 jkyle https://bugs.launchpad.net/fuel/+bug/1246030
19:59 dhblaz joined #fuel
20:01 xarses jkyle, have you checked http://docs.mirantis.com/fuel-dev/de​velop/env.html#building-the-fuel-iso
20:02 xarses An amount of the build process requires Debian/Ubuntu tools to build the Ubuntu sections of the ISO
20:03 xarses which might also explain the discrepancy in the function of the echo
20:05 dhblaz Here is one kernel panic:
20:05 dhblaz http://paste.openstack.org/​show/WvP7mLLAtguXDeh0xAnP/
20:12 jkyle I think it's more likely a diff in the build tools, both use gnu echo and 'echo -e' should work
20:12 jkyle but, yeah. I'll try it on ubuntu
20:15 jkyle ubuntu isos are wonky too, I can't mount them on my mac either
20:20 dhblaz Keep in mind that the shell echo command is different than /bin/echo
20:22 jkyle true
20:28 teran joined #fuel
21:06 e0ne joined #fuel
21:09 dhblaz xarses: did you see the kernel panic message I posted?
21:12 jsergent joined #fuel
21:13 mihgen joined #fuel
21:14 IlyaE joined #fuel
21:15 xarses Some one noted that line 1 references that there may be something in the "Integrated Management Log"
21:18 xarses hmm some one else mentioned that it's probably the hpwdt (watchdog)
21:19 xarses you might be able to blacklist the module https://wiki.archlinux.org/index.php/ker​nel_modules#Using_kernel_command_line_2
21:19 xarses modprobe.blacklist=hpwdt
21:20 xarses dhblaz: ^^
21:21 jsergent joined #fuel
21:22 dhblaz The IML shows:
21:22 dhblaz "POST error 1719 : A controller failure event occurred prior to this power-up"
21:22 jsergent joined #fuel
21:24 jsergent joined #fuel
21:30 xarses Do you have a smart array P410 controller ? http://h20566.www2.hp.com/portal/site/hpsc/templat​e.PAGE/public/kb/docDisplay/?sp4ts.oid=3883890&amp​;spf_p.tpst=kbDocDisplay&amp;spf_p.prp_kbDocDispla​y=wsrp-navigationalState%3DdocId%253Demr_na-c02910​003-4%257CdocLocale%253D%257CcalledBy%253D&amp;jav​ax.portlet.begCacheTok=com.vignette.cachetoken&amp​;javax.portlet.endCacheTok=com.vignette.cachetoken
21:31 Tingal joined #fuel
21:31 xarses there is also something similar for E200 controller http://h20566.www2.hp.com/portal/site/hpsc/templat​e.PAGE/public/kb/docDisplay/?sp4ts.oid=1157688&amp​;spf_p.tpst=kbDocDisplay&amp;spf_p.prp_kbDocDispla​y=wsrp-navigationalState%3DdocId%253Demr_na-c01917​689-1%257CdocLocale%253D%257CcalledBy%253D&amp;jav​ax.portlet.begCacheTok=com.vignette.cachetoken&amp​;javax.portlet.endCacheTok=com.vignette.cachetoken
21:40 dhblaz It is a P711
21:46 e0ne joined #fuel
21:58 dhblaz Andrew reports that we still get the kernel panic after modprobe.blacklist=hpwdt.  If he gets another stack trace I'll post it tomorrow
22:02 dhblaz I think our problem may be related to this:
22:02 dhblaz https://bugzilla.kernel.org/show_bug.cgi?id=60758
22:07 e0ne joined #fuel
22:49 e0ne joined #fuel
22:57 e0ne joined #fuel
23:28 e0ne joined #fuel
23:30 e0ne_ joined #fuel
23:58 e0ne joined #fuel
23:59 e0ne joined #fuel

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