15:00:59 <bswartz> #startmeeting manila
15:01:02 <openstack> Meeting started Thu Aug 14 15:00:59 2014 UTC and is due to finish in 60 minutes.  The chair is bswartz. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:03 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:05 <openstack> The meeting name has been set to 'manila'
15:01:07 <cknight> Hi
15:01:08 <bswartz> good morning folks
15:01:14 <vponomaryov> hi
15:01:20 <xyang1> hi
15:01:25 <shamail> Hello
15:01:31 <bswartz> #agenda https://wiki.openstack.org/wiki/Manila/Meetings
15:01:31 <csaba> hi
15:01:39 <rraja> hi
15:01:42 <vvechkanov> hi
15:01:50 <bswartz> I'm traveling today so I'm going to try to keep this quick
15:01:58 <bswartz> #topic incubation update
15:02:22 <bswartz> for those of you who weren't aware, the TC considered our incubation request on Tuesday
15:02:30 <bswartz> #link http://eavesdrop.openstack.org/meetings/tc/2014/tc.2014-08-12-20.02.log.html
15:02:42 <bswartz> start at like UTC 20:22
15:02:53 <bswartz> I'll let you all read the logs
15:02:55 <bswartz> any questions?
15:03:27 <vponomaryov> bswartz: there is expected second round of consideration?
15:03:54 <bswartz> yeah they should put us on the agenda again either next week or the week after
15:03:57 <bswartz> and give us a decision
15:04:11 <bswartz> we'll have to see
15:04:18 <bswartz> personally I thought the meeting went well
15:04:48 <shamail> It seems like good dialogue.
15:04:51 <bswartz> okay we can move on
15:05:11 <bswartz> #topic  Cirros status
15:05:29 <bswartz> csaba: what have you got?
15:06:49 <bswartz> uh oh did my network connection die?
15:06:55 <vponomaryov> no
15:07:00 <bswartz> whew
15:07:52 <bswartz> okay maybe we'll skip this topic and come back to it
15:07:55 <bswartz> #topic dev status
15:08:02 <vponomaryov> Dev status for last week:
15:08:15 <vponomaryov> 1) Generic driver is distro-independent now. It does not depend on ubuntu-specific stuff anymore. It uses only two libs: 'exportfs' and 'net' to handle shares, so Generic driver is able to work with any linux image, that has NFS and CIFS preinstalled and its mentioned share handlers.
15:08:15 <vponomaryov> gerrit: #link https://review.openstack.org/112279
15:08:15 <vponomaryov> status: merged
15:08:34 <rraja> bswartz: csaba got disconnected. he's trying to reconnect.
15:08:42 <vponomaryov> 2) Cmode driver refactor (removing dependency on 7mode driver)
15:08:42 <vponomaryov> bp: #link https://blueprints.launchpad.net/manila/+spec/remove-netapp-7mode-driver
15:08:42 <vponomaryov> gerrit: #link https://review.openstack.org/113596
15:08:42 <vponomaryov> status: ready for review, testing is in progress
15:08:50 <csaba> bswartz: I'm back, sorry guys
15:08:50 <bswartz> vponomaryov: how about samba 3.5 vs 4.0 -- any differences that we depend on?
15:09:00 <bswartz> csaba: we'll come back to your topic
15:09:12 <vponomaryov> bswartz: it is related to image
15:09:33 <vponomaryov> bswartz: Is your question about image size change?
15:09:34 <bswartz> I know that 4.0 is supposed to be dramatically smaller, therefore we want it
15:09:47 <bswartz> but do we depend on anything that changes between the 2 versions?
15:09:52 <vponomaryov> bswartz: I think csaba know more about it
15:10:10 <bswartz> does anyone know?
15:10:16 <bswartz> I suppose we just need to test it to find out
15:10:27 <csaba> bswartz: well in case of cirros it is expected to be an easy change
15:10:38 <bswartz> it would be ideal if the generic driver was compatible with both samba versions
15:10:41 <csaba> as buildroot has recipe for samba4
15:11:14 <vponomaryov> bswartz: I don't think there is incompatible change of configuration
15:11:19 <csaba> actually different recipe for samba and samba4, I'll have to check if they can deployed at the same time
15:11:32 <bswartz> no I don't think anyone would deploy both
15:11:50 <bswartz> the point is not to force users to use one if they'd rather use the other
15:12:13 <csaba> well you mean then to have two image flavors?
15:12:20 <vponomaryov> are we finished with status?
15:12:30 <bswartz> no the "official" image should eventually use samba4
15:12:41 <bswartz> but I expect admins to roll their own images somewhat frequently
15:13:00 <bswartz> yes ty vponomaryov
15:13:13 <bswartz> #topic cirros update
15:13:24 <bswartz> okay csaba
15:13:30 <csaba> bswartz: thx
15:14:08 <csaba> so I can just have negative reports, but sharpens the picture a bit..
15:14:19 <csaba> we had two progressive assumptions with vponomaryov
15:14:37 <csaba> a) the issue is that nfs-utils on cirros image is too old
15:14:41 <bswartz> #link https://github.com/csabahenk/cirros/issues/9
15:15:05 <csaba> b) the issue is that the kernel version and the nfs-utils version on the image are incompatibble
15:15:33 <csaba> (so if b), hypotetycally both up and downgrade can be a way out, with o/c upgrade preferred)
15:15:44 <bswartz> interesting
15:15:53 <bswartz> how did we end up with an old version?
15:16:00 <csaba> a) is clearly refuted by the fact that the ubuntu image has older nfs-utils than cirros
15:16:06 <bswartz> can't we use whatever version we want?
15:16:27 <csaba> so ubuntu has: linux 3.2.0, nfs-utils 1.2.5
15:16:45 <csaba> cirros has: linux 3.11, nfs-utils 1.2.6
15:17:10 <csaba> and actual stable versions are: linux 3.16, nfs-utils 1.3.0, FYI
15:17:15 <bswartz> I hope to have some time to look into this personally before next week
15:17:30 <csaba> so what I checked is cirros with 3.2.0 kernel
15:17:58 <csaba> whereby there is just one micro version difference from ubuntu wrt nfs-utils
15:18:04 <csaba> and that does not work either
15:18:27 <csaba> which lessens the probability of b) hypothesis working out
15:18:50 <csaba> however, upgrading everything is still something that we can and should do
15:18:58 <bswartz> what makes us think the problem is with the kernel and/or nfsutils?
15:19:20 <vponomaryov> bswartz: debug showed that the problem is in NFS daemon
15:19:31 <bswartz> yeah it makes sense -- I just want to know how we're sure
15:19:31 <csaba> bswartz: weird stuff in dmesg about nfs
15:19:33 <vponomaryov> bswartz: itis provided by nfs-utils lib in cirros
15:19:36 <bswartz> oh
15:19:52 <bswartz> okay I'll take your word for it until I can try it out on my own
15:20:08 <csaba> so what I think has a chance would be hypothesis c) -- uclibc issue
15:20:09 <bswartz> would more people looking at this help you csaba?
15:20:20 <csaba> bswartz: yes of course
15:20:25 <bswartz> hmmmm
15:20:31 <csaba> bswartz: I wanted to add a little info about that
15:20:37 <bswartz> yes uclibc could be a culprit
15:20:37 <csaba> how to get at it
15:20:46 <bswartz> we've tried to compile binaries for cirros before and run into that issue
15:20:48 <csaba> ie. you don't need openstack environment
15:21:09 <csaba> just need to fire up the image in say libvirt
15:21:14 <bswartz> right
15:21:28 <csaba> and start nfsd (/etc/init.d/S60nfsd)
15:21:33 <csaba> and try to mount from host
15:21:34 <bswartz> csaba: any special requirements to build your image?
15:21:52 <bswartz> or can I git clone ; make all
15:21:53 <bswartz> ?
15:22:20 <csaba> it's practical to edit the boot options to have dslist=nocloud in kernel cmdline... otherwise you have to wait a cloud-init timeout at boot
15:22:32 <bswartz> good to know
15:22:35 <csaba> bswartz: the readme elaborates
15:22:54 <bswartz> #link https://github.com/csabahenk/cirros/blob/manila-service-generic/README
15:22:58 <csaba> basically standard dev enviromentd
15:23:05 <bswartz> k
15:23:23 <csaba> recently I added ruby, which is used in the initrd-trimming srcipt
15:23:37 <bswartz> anything else csaba?
15:24:06 <csaba> (recall cirros' trick to reuse the same image as the rootfs for initrd... I have a ruby script that weeds out things that are not needed in early userspace)
15:24:25 <csaba> yes
15:24:51 <csaba> so my vision is to polish the bug reproduction until it can be submitted upsrtream
15:24:58 <csaba> (buildroot/ uClibc)
15:25:03 <bswartz> okay awesome
15:25:17 <bswartz> I'll ping you if  I discover anything on my own
15:25:24 <csaba> which involves syncing up buildroot with upstream and trying nfs-utils 1.3.0
15:25:42 <csaba> (nb. buildroot still has 1.2.6 so it has to be done manually)
15:25:50 <csaba> however, next week I'm on leave
15:26:00 <csaba> so I can pick these up the week after
15:26:02 <bswartz> ah
15:26:22 <bswartz> okay
15:26:28 <csaba> (that's all)
15:26:34 <bswartz> ok
15:26:37 <bswartz> #topic open discussion
15:27:09 <bswartz> anyone have anything?
15:27:26 <bswartz> vponomaryov: I have something
15:27:37 <vponomaryov> yes?
15:27:46 <bswartz> it seems likely that our dropbox files is NOT getting cached
15:27:53 <bswartz> and it's simply downloading extremely fast
15:28:13 <bswartz> I'm going to get together with the infra folks and they are likely to have a way to help us
15:28:30 <bswartz> they have mirrors and stuff
15:28:48 <vponomaryov> bswartz: great news
15:29:07 <bswartz> I've got some names
15:29:16 <bswartz> and I'll see if I can reach them today/tomorrow
15:29:32 <bswartz> anything else?
15:29:46 <vponomaryov> not from me
15:29:56 <bswartz> yay 30 minute meeting!
15:30:01 <bswartz> thanks all
15:30:04 <deepakcs> bye
15:30:08 <vvechkanov> bye
15:30:10 <bswartz> I'll send out more notes when I head from the TC again
15:30:13 <bswartz> hear*
15:30:16 <deepakcs> cool
15:30:21 <bswartz> #endmeeting