Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-07-30

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

All times shown according to UTC.

Time Nick Message
00:36 rmoe joined #fuel
00:36 geekinutah joined #fuel
01:13 KodiakF_athome Just a heads up on a plain installation of Virtualbox just running the launch.sh fails right out of the gate:
01:14 KodiakF_athome erm nm it didn't load the kernel modules...
01:14 KodiakF_athome rebooting
01:14 KodiakF_athome left #fuel
01:18 mattgriffin joined #fuel
01:46 KodiakF_athome joined #fuel
01:48 KodiakF_athome Question:  How do the CentOS 6.5 physical hosts get regular package updates?  Does the fuel server also act as a yum/spacewalk server in addition to cobbler?  Do the CentOS physical hosts need to get out to the upstream CentOS repos?
02:00 geekinutah KodiakF_athome: IIRC fuel does have it's own yum repo
02:01 geekinutah as long as nodes are running yum update, then you could just introduce the package into the repo on the fuel node
02:02 geekinutah I bet there's a better way to do it though, I'll defer to the experts here
03:00 CephFan1 joined #fuel
03:57 ArminderS joined #fuel
04:07 adanin joined #fuel
04:16 Longgeek joined #fuel
04:18 Longgeek joined #fuel
04:19 Longgeek joined #fuel
04:49 xarses joined #fuel
05:49 e0ne joined #fuel
06:29 e0ne joined #fuel
06:36 topochan joined #fuel
06:46 taj joined #fuel
06:46 taj joined #fuel
06:54 hyperbaba joined #fuel
07:04 pasquier-s joined #fuel
07:51 adanin joined #fuel
07:54 e0ne joined #fuel
07:55 topochan joined #fuel
07:57 al_ex11 joined #fuel
07:58 hyperbaba joined #fuel
08:08 bookwar joined #fuel
08:52 e0ne joined #fuel
09:16 artem_panchenko joined #fuel
09:31 Longgeek_ joined #fuel
09:37 adanin joined #fuel
09:43 gleam_ joined #fuel
09:44 aedocw_ joined #fuel
09:44 igalic_ joined #fuel
09:45 brain461 joined #fuel
09:45 brain461 left #fuel
09:45 jkirnosova joined #fuel
09:46 kat_pimenova joined #fuel
10:02 e0ne joined #fuel
10:37 brain461 joined #fuel
11:21 al_ex11 joined #fuel
11:41 TVR_ in fuel 5.0 .. the docker interface seems to be causing me issues.... is there a way to disable this and not use docker at all?
11:42 TVR_ even when I uncheck it as an interface .. it still gets called during the puppet runs and seems to cause the system to fail for the fuel provisioning system.
11:42 TVR_ I would like to try 5.0 or 5.01 but this makes that not possible for me
11:44 evg_ TVR_: do you mean docker0?
11:44 TVR_ yes
11:45 evg_ TVR_: It's a linux bridge, fuel 5.0 can't live without it.
11:45 evg_ TVR_: What's a problem with docker0?
11:46 TVR_ when I allow it to be installed, my fuel install completes, but when I set up a cluster (HA, neutron with gre) the network comes up as neutron with unknown so half the menu is missing (the menu for setting the CIDR for the internal instances)
11:47 TVR_ the assumption (wrong or right) is that the parts added (docker interface) is giving the issue, and as I will not be using containers, I would like to to parralel the 4.1 setups that presently work
11:48 TVR_ did that solve some issue I am unarare of by adding the docker0 interface?
11:49 TVR_ ok.. so it creates all the dm devices, and all that is needed now I take it
11:50 al_ex11 joined #fuel
11:51 TVR_ so is fuel now running in a container?
11:51 TVR_ first exposure to 5.0.. so bare with me please
11:52 evg_ TVR_: I don't think the docker0 is a problem. All fuel components work in lxc linux conteiners and docker0 connects them together.
11:52 TVR_ I have a very good handle on 4.1, but have some issues with network performance so wanted to see how well icehouse performes vs havana
11:52 evg_ TVR_: it's a bridge. See brctl show
11:53 TVR_ ok..
11:53 TVR_ so, should it use that interface for it's PXE int?
11:54 evg_ TVR_: you mean fuelmenu?
11:54 TVR_ yes please
11:55 TVR_ when setting up the fuel settings
11:55 TVR_ after initial install, but before puppet builds the fuel box
11:56 evg_ TVR_: yes
11:56 TVR_ ok..
12:00 evg_ TVR_: stop, no no. set phisical nic.
12:01 evg_ TVR_: http://docs.mirantis.com/openstack/fuel/fuel-5.0/user-guide.html#pxe-setup
12:01 TVR_ due to the issues I have, I am putting the miami site back to 4.1 and will do the boston site  (try fuel 5.0) as I have physical hands there... so I will be trying this once more but closer to 5 hours from now
12:03 TVR_ thx for the doc... I will look it over and use it
12:32 igalic_ joined #fuel
12:40 odyssey4me joined #fuel
13:00 geekinutah joined #fuel
13:18 jobewan joined #fuel
13:46 al_ex11 joined #fuel
13:55 al_ex11 joined #fuel
14:04 al_ex11 joined #fuel
14:18 al_ex12 joined #fuel
14:21 taj joined #fuel
14:32 mattgriffin joined #fuel
14:45 adanin joined #fuel
14:56 angdraug joined #fuel
15:45 ArminderS joined #fuel
15:49 al_ex12 joined #fuel
16:01 e0ne joined #fuel
16:03 ArminderS- joined #fuel
16:09 rmoe joined #fuel
16:28 taj joined #fuel
16:29 jaypipes joined #fuel
16:50 blahRus joined #fuel
17:07 adanin joined #fuel
17:11 adanin joined #fuel
17:19 tatyana joined #fuel
17:19 TVR_ is there an issue with HP servers and fuel 5.0?
17:19 aleksandr_null There are could be a problem with raid controllers but it depends.
17:20 TVR_ "can't load firmware file "bnx2/bnx2-mips-09-6.2.lb.fw"
17:21 TVR_ so, because of this, ethtool shows eth0 and eth1 as down (onboard nics) and therefore it cannot register with fuel
17:21 TVR_ it bootstraps to the ramdisk... but seems to drop the drivers from there and loose the nics
17:21 angdraug https://bugs.launchpad.net/fuel/+bug/1323354
17:21 angdraug that it?
17:21 TVR_ HP DL360 G6
17:22 angdraug looks like it was a problem and 5.0 and is fixed in 5.0.1
17:23 TVR_ ok.. so THAT also explains my earlier issues I accidentally was blaming docker for... the other site uses the R410 which is also affected.
17:23 TVR_ do we have a link to 5.0.1?
17:26 TVR_ I don't see it on the mirantis site... unless my eyes have been lost now like my sanity
17:27 ArminderS- nope, can't see myself too
17:27 ArminderS- your eyes are good and you are sane :)
17:29 TVR_ bold statement on my sanity.... any word or when it is expected to be released?
17:37 TVR_ did someone post a pre-release in here some time back for 5.0.1? I seem to remeber it... and am wondering if the actual 5.0.1 isn't released yet if I could try that...
17:37 angdraug joined #fuel
17:42 xarses joined #fuel
17:42 angdraug 5.0.1 isn't released yet, I'll see if I can scare up a prerelease build for you to try
17:42 angdraug long as you're fine with voiding your warranty :p
17:42 TVR_ heh.. no worries... the G6 fell out of warranty many a day ago
17:44 angdraug there's nightly 5.1 builds published here:
17:44 angdraug https://fuel-jenkins.mirantis.com/view/ISO/job/publish_fuel_community_iso/
17:44 brain461 joined #fuel
17:44 TVR_ iso's.. do you have images, or do i just roll one from it
17:44 angdraug 5.1 has a lot more changes than just bugfixes, but it has that firmware bug fixed
17:45 angdraug usb images are also there
17:46 rmoe TVR_: you can also fallback to 2.6 kernel bootstrap images: http://docs.mirantis.com/openstack/fuel/fuel-5.0/release-notes.html#known-issues-in-mirantis-openstack-5-0 under "Bootstrap kernel issues on certain hardware"
17:46 angdraug if you have 5.0 deployed, you can even try out our brand new upgrade capability
17:46 angdraug there's a download link for an upgrade tarball labeled UPGD next to each build
17:47 TVR_ cool
17:48 Longgeek joined #fuel
17:49 TVR_ seems everyone uses jenkins for deployments
17:49 TVR_ heh
17:50 angdraug here's an upcoming update for the docs about how to use these upgrade tarballs:
17:50 angdraug https://review.openstack.org/106895
17:50 angdraug there's a link in a comment from Fuel CI on that review to HTML docs built from that commit
17:51 angdraug real cutting edge stuff :D
17:56 tatyana_ joined #fuel
18:47 Longgeek joined #fuel
18:48 geekinutah joined #fuel
18:52 e0ne joined #fuel
18:56 adanin joined #fuel
19:37 e0ne joined #fuel
19:58 jseutter hm, I installed fuel but the pxe networking was not configured.  If I reinstall how do I get to the fuelmenu before the machine reboots, or if I keep the same install, how do I rebuild the docker containers so that pxeboot can be served on eth1?
20:05 e0ne joined #fuel
20:18 teknoprep so i have to remove all of those partitions that exist on current HDD"s
20:18 teknoprep so ceph will install
20:18 teknoprep anyone have a quick command to destroy Partition tables
20:18 teknoprep i was trying this using a script
20:19 teknoprep http://pastebin.com/YZ0mETQJ
20:21 ArminderS- in fuel 5.0 with ceph as backend for eph/vol/rados.gw, do we have possibility of putting quotas on object storage
20:21 ArminderS- i can see volume quota
20:22 ArminderS- does object quota has its own quotas?
20:25 e0ne joined #fuel
20:42 ArminderS none?
20:43 angdraug Arminder: https://www.mail-archive.com/ceph-users@lists.ceph.com/msg08652.html
20:44 angdraug radosgw has quotas, but you have to set them up by hand
20:44 angdraug based on the rest of the thread, Dumpling version of radosgw which is what we had in 5.0 might be buggy
20:45 angdraug ...might have buggy quotas implementations
20:56 ArminderS thanks angdraug, will explore more of it
20:56 ArminderS i wonder how its going to react on quota full
21:03 teknoprep (/Stage[main]/Ceph::Osd/Exec[ceph-deploy osd prepare]/returns) change from notrun to 0 failed: ceph-deploy osd prepare node-210:/dev/sdb4 returned 1 instead of one of [0]
21:03 teknoprep sucks
21:05 rmoe you should just be able to wipe your osd partitions (dd if=/dev/zero of=/dev/sdb4 bs=1M count=10) and re-run puppet
21:05 teknoprep yeah i get it
21:05 teknoprep i used wipefs -a /dev/sda
21:05 teknoprep i dunno if that will work though
21:05 teknoprep even though wipefs should wipe all partitions
21:06 teknoprep how do i re-run puppet ?
21:06 rmoe what does parted /dev/sda print show?
21:06 teknoprep parted is not installed on bootstrap
21:06 rmoe (or whatever device your osd is using)
21:07 rmoe oh, if it's in bootstrap there won't be much you can do
21:07 rmoe just have to let it fail, clean up the partitions and re-run puppet
21:07 teknoprep ok
21:07 teknoprep how do i re-run puppet
21:07 rmoe you can either have fuel trigger the redeploy or run puppet apply /etc/puppet/manifests/site,pp on each node
21:08 teknoprep ok
21:08 teknoprep so how do i use parted... as parted /dev/sdb does not show much info.. just alist of commands
21:09 rmoe parted /dev/sdb print
21:09 teknoprep i am now in the OS.. after it errored
21:09 rmoe it'll show the partitions and the detected file system
21:09 teknoprep ok
21:09 rmoe you'll probably see it show xfs for sdb4
21:09 teknoprep yeah
21:09 teknoprep should i run that dd command ?
21:09 rmoe which is why ceph-deploy thinks it's an existing osd
21:09 teknoprep gotcha
21:10 teknoprep even after running dd if=/dev/zero of=/dev/sdb4 bs=1M count=10 --- the partition still shows up
21:10 teknoprep 4      445MB   1000GB  1000GB  xfs          primary
21:11 teknoprep rmoe ? any advice please ?
21:12 rmoe that's odd, can you use paste and show the full parted print command, followed by dd, followed again by parted?
21:12 teknoprep sure
21:13 teknoprep http://pastebin.com/yrnuzcZU
21:16 angdraug joined #fuel
21:16 rmoe here's what I get: http://paste.openstack.org/show/89250/
21:16 rmoe what kind of hardware do you have?
21:17 teknoprep supermicro... intel hardware... sata disk drives
21:17 teknoprep everything is server grade supermicro quad node systems
21:18 rmoe try wipefs with the same before/after parted commands?
21:18 teknoprep wait
21:18 teknoprep wwhen i am in parted
21:18 teknoprep it gives different output
21:18 teknoprep one sec
21:18 teknoprep nvm
21:19 teknoprep wipefs -a /dev/sdb4 ?
21:19 rmoe yeah, give that a shot
21:20 teknoprep lol nothing
21:20 teknoprep still shows up
21:20 teknoprep with xfs
21:20 teknoprep can i use parted to delete the partition ?
21:21 teknoprep (parted) print
21:21 teknoprep Error: /dev/sdb4: unrecognised disk label
21:22 teknoprep i wonder if its using a chached copy of partition after mount
21:22 teknoprep but after we did that it deleted it
21:22 teknoprep inside parted it has no info
21:22 rmoe is it moutned?
21:22 teknoprep doesn't seem to be under mounted
21:22 teknoprep mount
21:22 teknoprep how do you mount XFS??
21:23 adanin joined #fuel
21:23 rmoe mount /dev/vdb4 /mnt/ will work
21:23 rmoe if it has a valid filesytem
21:24 teknoprep have to specify fs
21:24 teknoprep wrong fs
21:24 teknoprep when i did -t xfs
21:24 teknoprep but it shouldnt need that
21:49 thehybridtech joined #fuel
22:14 rmoe so there's really no fs there and you should be good to run puppet again
23:50 mattgriffin joined #fuel

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