*** rcallawa has quit IRC | 00:01 | |
*** rcallawa has joined #openstack-manila | 00:02 | |
*** ho has quit IRC | 00:04 | |
*** ho has joined #openstack-manila | 00:06 | |
*** scottda has joined #openstack-manila | 00:18 | |
*** scottda has quit IRC | 00:18 | |
*** scottda has joined #openstack-manila | 00:20 | |
*** scottda has quit IRC | 00:22 | |
*** scottda has joined #openstack-manila | 00:24 | |
*** scottda has quit IRC | 00:39 | |
*** scottda has joined #openstack-manila | 00:43 | |
*** dalgaaf has quit IRC | 00:47 | |
*** scottda has quit IRC | 00:48 | |
*** akerr has quit IRC | 00:49 | |
*** rushil has quit IRC | 00:51 | |
*** scottda has joined #openstack-manila | 00:55 | |
*** scottda has quit IRC | 01:07 | |
*** scottda has joined #openstack-manila | 01:08 | |
*** scottda has quit IRC | 01:28 | |
*** scottda has joined #openstack-manila | 01:31 | |
*** cknight has joined #openstack-manila | 01:42 | |
*** scottda has quit IRC | 01:49 | |
*** cknight has quit IRC | 01:51 | |
*** cknight has joined #openstack-manila | 01:51 | |
*** kaisers1 has joined #openstack-manila | 02:31 | |
*** kaisers has quit IRC | 02:32 | |
*** tbarron1 has quit IRC | 02:39 | |
*** xyang1 has joined #openstack-manila | 03:08 | |
*** scottda has joined #openstack-manila | 03:31 | |
*** scottda has quit IRC | 03:32 | |
openstackgerrit | Merged openstack/python-manilaclient: Print share export locations more pretty https://review.openstack.org/165923 | 03:34 |
---|---|---|
openstackgerrit | Merged openstack/python-manilaclient: Make public shares visible for all tenants https://review.openstack.org/148869 | 03:52 |
*** cknight has quit IRC | 03:59 | |
*** ho has quit IRC | 04:08 | |
openstackgerrit | Merged openstack/manila: Imported Translations from Transifex https://review.openstack.org/165306 | 04:10 |
openstackgerrit | Merged openstack/manila: Enable bashate during pep8 run https://review.openstack.org/146915 | 04:16 |
*** jasonsb has joined #openstack-manila | 04:19 | |
*** jasonsb has quit IRC | 04:20 | |
*** jasonsb has joined #openstack-manila | 04:20 | |
*** scottda has joined #openstack-manila | 04:21 | |
*** scottda has quit IRC | 04:41 | |
*** scottda has joined #openstack-manila | 04:42 | |
*** chlong has quit IRC | 05:02 | |
*** scottda has quit IRC | 05:10 | |
*** chlong has joined #openstack-manila | 05:19 | |
*** rhefner has joined #openstack-manila | 05:35 | |
*** rhefner has left #openstack-manila | 05:40 | |
*** rhefner has joined #openstack-manila | 05:40 | |
*** sks has joined #openstack-manila | 05:54 | |
*** xyang1 has quit IRC | 06:00 | |
*** chlong has quit IRC | 06:01 | |
*** chlong has joined #openstack-manila | 06:13 | |
*** lpetrut has joined #openstack-manila | 06:36 | |
*** deepakcs has joined #openstack-manila | 06:40 | |
deepakcs | vponomaryov, ping | 06:42 |
deepakcs | vponomaryov, in latest devstack , i see sched error when manila configured with glusterfs_native driver (No valid host was found). | 06:43 |
deepakcs | vponomaryov, I never saw this before (atleast 2 months back) :) | 06:43 |
deepakcs | vponomaryov, so wonderign if anythign changed in the filter_sched (default) that may cause this to happen > | 06:43 |
deepakcs | vponomaryov, I looked at the stats reported by glsuter_native driver, they are similar like how generic reports (as far as fields and values are concerned) | 06:44 |
*** nkrinner has joined #openstack-manila | 07:20 | |
*** chlong has quit IRC | 07:45 | |
*** TobiasE has joined #openstack-manila | 07:48 | |
*** vbellur has quit IRC | 08:09 | |
*** tbarron has joined #openstack-manila | 08:18 | |
*** openstackgerrit has quit IRC | 08:22 | |
*** openstackgerrit has joined #openstack-manila | 08:22 | |
*** rraja has joined #openstack-manila | 08:23 | |
*** tbarron1 has joined #openstack-manila | 08:28 | |
*** deepakcs has quit IRC | 08:30 | |
*** tbarron has quit IRC | 08:31 | |
*** vbellur has joined #openstack-manila | 08:31 | |
*** rraja has quit IRC | 08:35 | |
*** rraja has joined #openstack-manila | 08:38 | |
*** lpetrut has quit IRC | 08:59 | |
*** ajo has joined #openstack-manila | 09:20 | |
*** HoloIRCUser7 has joined #openstack-manila | 09:24 | |
vponomaryov | deepakcs: answer is simple | 09:25 |
vponomaryov | deepakcs: share_type is used now always | 09:25 |
vponomaryov | deepakcs: and default share type has required key "driver_handles_share_servers" set to "True" | 09:26 |
vponomaryov | deepakcs: that is opposite to gluster drivers | 09:26 |
vponomaryov | deepakcs: either use your own share type, or set it via localrc for default one | 09:27 |
*** rraja has quit IRC | 09:33 | |
*** HoloIRCUser7 has quit IRC | 09:39 | |
vponomaryov | csaba: hello, please, inform deepakcs about answer for his question if possible ^ | 09:44 |
*** lpetrut has joined #openstack-manila | 10:20 | |
*** Maike has joined #openstack-manila | 10:30 | |
*** Maike_ has joined #openstack-manila | 10:31 | |
*** Maike has quit IRC | 10:35 | |
*** tbarron1 has quit IRC | 10:40 | |
*** erlon has joined #openstack-manila | 10:42 | |
*** rraja has joined #openstack-manila | 10:47 | |
csaba | vponomaryov: thx, I'll talk with deepakcs | 10:49 |
*** vbellur has quit IRC | 10:51 | |
*** tbarron has joined #openstack-manila | 10:59 | |
*** rraja_ has joined #openstack-manila | 11:00 | |
*** rraja has quit IRC | 11:01 | |
*** tbarron1 has joined #openstack-manila | 11:01 | |
*** tbarron has quit IRC | 11:04 | |
*** vbellur has joined #openstack-manila | 11:04 | |
*** rcallawa has quit IRC | 11:12 | |
*** timcl has joined #openstack-manila | 11:38 | |
*** ajo has quit IRC | 11:42 | |
*** ajo has joined #openstack-manila | 11:43 | |
*** ajo has quit IRC | 11:44 | |
*** ajo has joined #openstack-manila | 11:44 | |
*** ganso_ has joined #openstack-manila | 11:45 | |
openstackgerrit | Julia Varlamova proposed openstack/manila: Automatic cleanup of share_servers https://review.openstack.org/166182 | 11:51 |
*** rraja_ has quit IRC | 11:55 | |
*** sgotliv has quit IRC | 11:59 | |
*** eharney has quit IRC | 12:07 | |
*** porrua has joined #openstack-manila | 12:17 | |
*** vbellur has quit IRC | 12:32 | |
*** bswartz has quit IRC | 12:34 | |
*** tbarron1 has quit IRC | 12:35 | |
*** ajo has quit IRC | 12:43 | |
*** ajo has joined #openstack-manila | 12:43 | |
*** vbellur has joined #openstack-manila | 12:47 | |
*** akerr has joined #openstack-manila | 12:52 | |
*** eharney has joined #openstack-manila | 13:15 | |
*** Maike_ has quit IRC | 13:16 | |
*** xyang1 has joined #openstack-manila | 13:16 | |
*** Maike_ has joined #openstack-manila | 13:16 | |
*** timcl has quit IRC | 13:17 | |
*** xyang_ has joined #openstack-manila | 13:18 | |
*** bswartz has joined #openstack-manila | 13:19 | |
*** dustins has joined #openstack-manila | 13:22 | |
*** eharney has quit IRC | 13:25 | |
*** eharney has joined #openstack-manila | 13:26 | |
*** timcl has joined #openstack-manila | 13:30 | |
*** cknight has joined #openstack-manila | 13:30 | |
*** eharney has quit IRC | 13:33 | |
*** scottda has joined #openstack-manila | 13:44 | |
*** scottda has quit IRC | 13:48 | |
*** eharney has joined #openstack-manila | 13:49 | |
*** mtanino has joined #openstack-manila | 13:50 | |
vponomaryov | bswartz, xyang_: Hello, please, look at this client change - https://review.openstack.org/158704 It is the last one left, after this one up we can make tag of manilaclient | 13:54 |
vponomaryov | tag will be useful when we will set our manilaclient as dependency | 13:54 |
vponomaryov | and this tag will be minimal version | 13:54 |
*** cknight has quit IRC | 14:00 | |
xyang_ | vponomaryov: hi, sure | 14:06 |
*** esker has joined #openstack-manila | 14:06 | |
*** Maike_ has quit IRC | 14:08 | |
*** dgonzalez has joined #openstack-manila | 14:09 | |
*** cknight has joined #openstack-manila | 14:11 | |
mkoderer | vponomaryov: bswartz: just a quick comment to bug https://bugs.launchpad.net/manila/+bug/1434039 | 14:12 |
openstack | Launchpad bug 1434039 in Manila "Generic driver snapshot produces corrupt filesystems" [Undecided,New] | 14:12 |
mkoderer | IMHO I would disable the snapshot functionality in the generic drivers as long this is not fixed | 14:12 |
mkoderer | data loss is the worst case for customers :) | 14:12 |
bswartz | mkoderer: I think the problem will depend on which cinder backend you use behind manila generic driver | 14:13 |
bswartz | not all cinder drivers have the same snapshot semantics | 14:13 |
mkoderer | bswartz: are there backend supporting live snapshots? | 14:13 |
bswartz | netapp backends do | 14:13 |
mkoderer | bswartz: this needs to get documented | 14:14 |
bswartz | although I haven't personally tested manila generic on top of netapp cinder | 14:14 |
bswartz | mkoderer: is it possible to detect this bug with an automated scenario test? | 14:14 |
mkoderer | bswartz: who will ever do that.. using generic with netapp ;) | 14:14 |
bswartz | mkoderer: I know -- that's why it's not tested | 14:14 |
mkoderer | bswartz: it's maybe more a stress test than a scenario test | 14:15 |
*** esker has quit IRC | 14:15 | |
bswartz | if we have a test that takes a snapshot and clones it and tries to mount it then we can check all driers for this bad behavior | 14:15 |
bswartz | drivers* | 14:15 |
mkoderer | bswartz: since we have to create a certain load to reporduce it | 14:15 |
*** esker has joined #openstack-manila | 14:15 | |
bswartz | oh so it's not 100% of the time? | 14:15 |
mkoderer | bswartz: no I don't have any numbers | 14:16 |
bswartz | could we make it less likely by doing sync operations on the service VM at the time of the snapshot? | 14:16 |
bswartz | but my question is, are there times when it actually works okay? | 14:16 |
mkoderer | bswartz: I think the issue can be solved in the gerneric driver with "sync" or "unmount" | 14:17 |
bswartz | vponomaryov: can we ssh to the service VM and "sudo sync"? | 14:17 |
*** tbarron has joined #openstack-manila | 14:17 | |
mkoderer | but this will lead to a (short) downtime for NFS clients | 14:17 |
bswartz | sync will not cause downtime | 14:18 |
bswartz | umount will | 14:18 |
mkoderer | bswartz: you would need to sync+freeze if you would like to do it in the clean way | 14:19 |
*** nkrinner has quit IRC | 14:20 | |
*** vbellur has quit IRC | 14:22 | |
vponomaryov | bswartz, mkoderer: how are you gonna make sure some user does not write data there? | 14:22 |
vponomaryov | mkoderer, bswartz: while syncing? | 14:22 |
mkoderer | vponomaryov: that's why "freeze" the vm is necessary | 14:24 |
mkoderer | = downtime | 14:24 |
vponomaryov | mkoderer: then what did bswartz mean by "sync will not cause downtime" ? | 14:25 |
vponomaryov | bswartz? | 14:25 |
vponomaryov | bswartz: "vponomaryov: can we ssh to the service VM and "sudo sync"?" we can ssh, yes. | 14:26 |
*** rushil has joined #openstack-manila | 14:31 | |
mkoderer | bswartz: to be honest I have doubts that the cinder netapp driver support live snapshots | 14:40 |
vponomaryov | mkoderer: backend itself does it | 14:40 |
mkoderer | vponomaryov: manila backend? | 14:40 |
bswartz | sorry I stepped away from my desk for a few minutes | 14:40 |
vponomaryov | mkoderer: netapp storage | 14:41 |
bswartz | mkoderer: why do you have doubts about the cinder driver? | 14:41 |
bswartz | a VM freeze would allow us to pause I/O long enough to get a clean snapshot | 14:41 |
bswartz | does nova provide that? | 14:41 |
mkoderer | bswartz: because it's a block device.. how do you know anyting about the underling FS | 14:41 |
*** dustins_ has joined #openstack-manila | 14:42 | |
bswartz | mkoderer: a netapp snapshot of a block device is guaranteed to be "crash consistent" | 14:42 |
*** rcallawa has joined #openstack-manila | 14:42 | |
bswartz | that means that any writes to the device will appear to have been in-order | 14:42 |
bswartz | just like if you pulled the plug on your fileserver | 14:42 |
mkoderer | bswartz: but what about client fs caches? | 14:43 |
bswartz | you can get a more consistent snapshot by quiecsing the client and flushing caches | 14:43 |
bswartz | you're correct that if the client cached writes and you take a snapshot, the writes will no be reflected | 14:44 |
bswartz | however what is reflected will still be consistent | 14:44 |
bswartz | it will just be a few seconds old | 14:44 |
*** dustins has quit IRC | 14:44 | |
bswartz | I still think it's important to point out the difference between downtime caused by an unmount of the filesystem vs a pause in I/O caused by a VM freeze | 14:46 |
bswartz | in one case the manila client will see I/O errors, and in the other case, he will just see silence for a few seconds | 14:46 |
bswartz | silence is a lot better than errors, because nearly every filesystem client can deal with network hiccups by issuing retries and having timesouts, etc | 14:47 |
mkoderer | bswartz: yep I agree | 14:48 |
mkoderer | bswartz: better having some reconnections ongoing than data loss | 14:48 |
*** sks has quit IRC | 14:48 | |
bswartz | so we don't really want to freeze the whole VM, just the frontend network | 14:48 |
bswartz | mkoderer: I have targeted the bug to liberty because my gut feeling is that this would be an expensive/complex fix | 14:54 |
bswartz | if someone can show a low risk/simple fix for this then we can consider pulling it into kilo | 14:54 |
bswartz | I don't think we would block the release for this though because it's not a new problem and fixing it requires adding new functionality | 14:55 |
*** juzuluag has joined #openstack-manila | 14:59 | |
*** porrua has quit IRC | 15:10 | |
*** TobiasE has left #openstack-manila | 15:11 | |
*** akerr has quit IRC | 15:12 | |
*** akerr_ has joined #openstack-manila | 15:12 | |
*** timcl has quit IRC | 15:18 | |
*** openstackgerrit has quit IRC | 15:21 | |
*** openstackgerrit has joined #openstack-manila | 15:22 | |
*** rushil_ has joined #openstack-manila | 15:23 | |
openstackgerrit | Merged openstack/python-manilaclient: Implement private share_types https://review.openstack.org/158704 | 15:24 |
*** rushil has quit IRC | 15:26 | |
*** dustins_ has quit IRC | 15:37 | |
*** jasonsb has quit IRC | 15:38 | |
*** scottda has joined #openstack-manila | 15:44 | |
*** scottda has quit IRC | 15:45 | |
*** timcl has joined #openstack-manila | 15:48 | |
*** scottda has joined #openstack-manila | 15:49 | |
*** scottda has quit IRC | 15:50 | |
*** rcallawa has quit IRC | 15:51 | |
*** rcallawa has joined #openstack-manila | 15:54 | |
*** scottda has joined #openstack-manila | 16:05 | |
*** scottda has quit IRC | 16:07 | |
*** scottda has joined #openstack-manila | 16:08 | |
*** scottda has quit IRC | 16:12 | |
*** esker has quit IRC | 16:15 | |
vponomaryov | bswartz: who is able to do tagging of manilaclient? | 16:20 |
bswartz | vponomaryov: http://git.openstack.org/cgit/openstack/python-manilaclient/tag/?id=1.0.3 | 16:43 |
*** timcl has quit IRC | 16:43 | |
vponomaryov | bswartz: thanks =) | 16:45 |
*** juzuluag has quit IRC | 16:48 | |
*** tbarron has quit IRC | 16:49 | |
*** dustins has joined #openstack-manila | 16:54 | |
*** tbarron has joined #openstack-manila | 16:58 | |
*** rcallawa_ has joined #openstack-manila | 17:01 | |
*** rcallawa has quit IRC | 17:04 | |
*** zaitcev has joined #openstack-manila | 17:17 | |
*** vbellur has joined #openstack-manila | 17:21 | |
*** openstackgerrit has quit IRC | 17:21 | |
*** porrua has joined #openstack-manila | 17:21 | |
*** openstackgerrit has joined #openstack-manila | 17:22 | |
*** Maike_ has joined #openstack-manila | 17:31 | |
*** Maike__ has joined #openstack-manila | 17:33 | |
*** Maike_ has quit IRC | 17:36 | |
*** vbellur has quit IRC | 17:40 | |
*** vbellur has joined #openstack-manila | 17:42 | |
*** vbellur has quit IRC | 18:00 | |
Maike__ | Is that new functionality that a share_service_network is created during the stack.sh? | 18:55 |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add share state verification for API 'unmanage' https://review.openstack.org/166348 | 19:03 |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add share state verification for API 'unmanage' https://review.openstack.org/166348 | 19:14 |
vponomaryov | Maike__: devstack is able to create one Nova instance per each generic driver enabled to not handle share servers | 19:19 |
vponomaryov | Maike__: generic driver uses Cinder as backend, and Cinder volume can be handled only by means of Nova API and Cinder API. So, Nova VM is requried in any case for Generic driver. | 19:20 |
vponomaryov | Maike__: by default it is not installed | 19:21 |
Maike__ | vponomaryov: ok. I will check my configuration! At the moment there is no valid host found and no action in cinder logs during the manila create | 19:22 |
vponomaryov | Maike__: 99% that scheduler filtered hosts with inappropriate driver mode | 19:23 |
vponomaryov | look at "manila type-list" | 19:23 |
vponomaryov | one of share types was used | 19:24 |
vponomaryov | if you did not provide it explicitly - default one will be used | 19:24 |
Maike__ | for me there is just the default one | 19:24 |
vponomaryov | and looks like you do not have hosts with driver mode that share type requires | 19:24 |
vponomaryov | just change "driver_handles_share_servers" to vice versa | 19:25 |
vponomaryov | for this the only share type | 19:25 |
Maike__ | I've set driver_handles_share_servers = True | 19:26 |
vponomaryov | Maike__: but do you have such backend configured? | 19:27 |
vponomaryov | Maike__: look at /etc/manila/manila.conf | 19:27 |
vponomaryov | Maike__: ther eshould be same opt for each driver | 19:27 |
Maike__ | vponomaryov: yes, it's configured in /etc/manila/manila.conf! anywhere else? | 19:29 |
vponomaryov | Maike__: show me your manila.conf, "manila extra-specs-list" and "share creation" command | 19:30 |
Maike__ | vponomaryov: http://paste.openstack.org/show/194021/ | 19:34 |
vponomaryov | Maike__: ah =) | 19:34 |
vponomaryov | Maike__: key : 1 -> this is the reason | 19:35 |
*** rcallawa_ has quit IRC | 19:35 | |
vponomaryov | Maike__: just remove it | 19:35 |
Maike__ | vponomaryov: I can fix this with type-key unset key ? | 19:35 |
vponomaryov | Maike__: right | 19:35 |
vponomaryov | Maike__: scheduler looks for this key on hosts, does not found it and returns to you "No hosts found" | 19:36 |
vponomaryov | Maike__: it is expected as far, this is random key | 19:36 |
*** rcallawa has joined #openstack-manila | 19:36 | |
Maike__ | vponomaryov: oh, ok. and in which case can I use this ? | 19:37 |
vponomaryov | Maike__: what this? | 19:37 |
Maike__ | sorry.. this key=1 | 19:37 |
vponomaryov | Maike__: I guess you set it playing with devstack's feature "MANILA_CONFIGOPT" | 19:38 |
vponomaryov | Maike__: if not manually | 19:38 |
vponomaryov | sorry, this one - "MANILA_DEFAULT_SHARE_TYPE_EXTRA_SPECS" | 19:38 |
Maike__ | I think I set it accidentally manually ;) | 19:38 |
Maike__ | yes | 19:38 |
Maike__ | !! | 19:38 |
openstack | Maike__: Error: "!" is not a valid command. | 19:38 |
vponomaryov | Maike__: manila type-key default unset key | 19:39 |
Maike__ | vponomaryov: now it works | 19:39 |
Maike__ | vponomaryov: thanks!! I played around with MANILA_DEFAULT_SHARE_TYPE_EXTRA_SPECS because of an error during stack.sh | 19:40 |
vponomaryov | Maike__: sorry, I need to go already | 19:40 |
vponomaryov | bye | 19:40 |
Maike__ | vponomaryov: bye | 19:41 |
*** rcallawa has quit IRC | 19:41 | |
Maike__ | vponomaryov: but it doesn't work.. same error message | 19:41 |
vponomaryov | try set "driver_handles_share_servers" to "True" instead of "1" | 19:42 |
vponomaryov | for share-type | 19:43 |
vponomaryov | Maike__: "manila type-key default set driver_handles_share_servers=True" | 19:43 |
*** lpetrut has quit IRC | 19:44 | |
Maike__ | vponomaryov: doesn't work :( | 19:45 |
*** jasonsb has joined #openstack-manila | 20:11 | |
*** Maike__ has quit IRC | 20:14 | |
*** erlon is now known as erlon_away | 20:21 | |
*** lpetrut has joined #openstack-manila | 20:24 | |
*** rcallawa has joined #openstack-manila | 20:31 | |
*** eharney has quit IRC | 20:38 | |
*** dustins has quit IRC | 20:39 | |
*** jasonsb has quit IRC | 20:40 | |
*** bswartz has quit IRC | 20:40 | |
*** lpetrut has quit IRC | 20:44 | |
*** porrua has quit IRC | 21:00 | |
*** esker has joined #openstack-manila | 21:03 | |
*** akerr_ has quit IRC | 21:07 | |
*** lpetrut has joined #openstack-manila | 21:07 | |
*** rcallawa has quit IRC | 21:08 | |
*** bswartz has joined #openstack-manila | 21:15 | |
*** esker has quit IRC | 21:38 | |
*** scottda has joined #openstack-manila | 21:46 | |
*** lpetrut has quit IRC | 21:49 | |
*** tbarron has quit IRC | 21:59 | |
*** tbarron has joined #openstack-manila | 22:00 | |
*** tbarron has quit IRC | 22:04 | |
*** scottda has quit IRC | 22:05 | |
*** esker has joined #openstack-manila | 22:15 | |
*** esker has quit IRC | 22:19 | |
*** esker has joined #openstack-manila | 22:19 | |
*** scottda has joined #openstack-manila | 22:21 | |
*** cknight has quit IRC | 22:39 | |
*** erlon_away has quit IRC | 22:41 | |
*** scottda has quit IRC | 22:52 | |
*** tbarron has joined #openstack-manila | 22:54 | |
*** scottda has joined #openstack-manila | 22:54 | |
*** ganso_ has quit IRC | 22:57 | |
*** scottda has quit IRC | 23:00 | |
*** scottda has joined #openstack-manila | 23:01 | |
*** mtanino has quit IRC | 23:08 | |
*** esker has quit IRC | 23:15 | |
*** esker has joined #openstack-manila | 23:18 | |
*** esker has quit IRC | 23:38 | |
*** akerr has joined #openstack-manila | 23:42 | |
*** scottda has quit IRC | 23:44 | |
*** akerr_ has joined #openstack-manila | 23:46 | |
*** akerr has quit IRC | 23:49 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!