| *** merooney has quit IRC | 00:01 | |
| *** xyang has joined #openstack-manila | 00:05 | |
| *** xyang has quit IRC | 00:06 | |
| *** yangyapeng has quit IRC | 00:15 | |
| *** Suyash has quit IRC | 00:45 | |
| *** yangyapeng has joined #openstack-manila | 01:19 | |
| *** xyang has joined #openstack-manila | 01:28 | |
| *** xyang has quit IRC | 01:31 | |
| *** Suyash has joined #openstack-manila | 01:32 | |
| *** draynium has quit IRC | 01:33 | |
| zhongjun_ | gouthamr_: ping | 01:46 |
|---|---|---|
| gouthamr_ | zhongjun_: pong | 01:46 |
| zhongjun_ | gouthamr_: Does the replication need to change something when I change the share to multiple protocols share? | 01:47 |
| gouthamr_ | zhongjun_: i can only think of access_rules and export_locations.. both of these behave the same as they do with shares today.. | 01:50 |
| zhongjun_ | gouthamr_: share replication feature about framwork part. | 01:50 |
| gouthamr_ | zhongjun_: i can't think of any differences.. today share replication works the same for NFS/CIFS/any other protocol | 01:51 |
| zhongjun_ | gouthamr_: I also think so, but still need test. | 01:54 |
| gouthamr_ | zhongjun_: +1 | 01:54 |
| *** gouthamr_ is now known as gouthamr | 01:54 | |
| *** baojg has joined #openstack-manila | 01:54 | |
| zhongjun_ | Do you think this method is feasible :" just add new NFS_CIFS protocols support when we create a share." | 01:55 |
| zhongjun_ | gouthamr: If this method don't have big holes, I think I need to do a demo to test it. | 01:57 |
| gouthamr | zhongjun_: We need to think about how the following things will work: export paths, access rules, security services... | 02:01 |
| gouthamr | that's the part i find confusing as of now, looking at how the NetApp driver is implemented.. | 02:01 |
| zhongjun_ | gouthamr: export paths will export with protocol. please see the examples in the doc(https://review.openstack.org/#/c/329392 ). Is this suitable for NetApp drivers? | 02:07 |
| gouthamr | zhongjun_: thanks.. will review that spec :) | 02:13 |
| zhongjun_ | gouthamr: The security services as the same as before. We could add multiple security services in one share network. then create a share with one share network. The driver will do "setup_server" function to setup a share service(config AD or LDAP domain), then do "create_share" function to decide whether is a suitable security service for share. That is my understanding. so the security services method not need to change. It is up to | 02:17 |
| zhongjun_ | driver | 02:17 |
| zhongjun_ | gouthamr: thanks gouthamr | 02:18 |
| zhongjun_ | gouthamr: :) | 02:18 |
| gouthamr | zhongjun_: nice, you're welcome.. | 02:18 |
| *** houming has joined #openstack-manila | 02:29 | |
| *** gouthamr has quit IRC | 02:55 | |
| *** bswartz has joined #openstack-manila | 02:56 | |
| *** cknight has quit IRC | 02:59 | |
| *** cknight has joined #openstack-manila | 02:59 | |
| *** cknight has quit IRC | 03:27 | |
| *** cknight has joined #openstack-manila | 03:27 | |
| *** baojg has quit IRC | 03:50 | |
| *** baojg has joined #openstack-manila | 03:52 | |
| *** cknight1 has joined #openstack-manila | 04:02 | |
| *** cknight has quit IRC | 04:04 | |
| *** lpetrut has joined #openstack-manila | 04:32 | |
| *** Suyash has quit IRC | 04:51 | |
| *** lpetrut has quit IRC | 04:55 | |
| *** houming has quit IRC | 05:05 | |
| *** nkrinner_afk is now known as nkrinner | 05:17 | |
| *** lpetrut has joined #openstack-manila | 05:23 | |
| *** cknight1 has quit IRC | 05:24 | |
| *** shausy has joined #openstack-manila | 05:27 | |
| *** lpetrut has quit IRC | 05:28 | |
| *** houming has joined #openstack-manila | 05:43 | |
| *** houming has quit IRC | 05:49 | |
| *** baojg has quit IRC | 05:56 | |
| *** lpetrut has joined #openstack-manila | 06:11 | |
| *** baojg has joined #openstack-manila | 06:18 | |
| *** baojg_ has joined #openstack-manila | 06:22 | |
| *** baojg has quit IRC | 06:23 | |
| *** baojg_ has quit IRC | 06:31 | |
| *** baojg has joined #openstack-manila | 06:35 | |
| *** baojg has quit IRC | 06:54 | |
| *** lpetrut has quit IRC | 06:55 | |
| *** baojg has joined #openstack-manila | 07:07 | |
| *** baojg has quit IRC | 07:23 | |
| openstackgerrit | zhaohua proposed openstack/manila: Huawei driver support access of all IPs https://review.openstack.org/311052 | 07:27 |
| openstackgerrit | jay.xu proposed openstack/manila: Add gateway in network_info and share network API https://review.openstack.org/306990 | 07:27 |
| openstackgerrit | jay.xu proposed openstack/manila: Add gateway in network_info and share network API https://review.openstack.org/306990 | 08:01 |
| *** dmk0202 has joined #openstack-manila | 08:17 | |
| *** baojg has joined #openstack-manila | 08:35 | |
| *** rraja has joined #openstack-manila | 08:50 | |
| *** jcsp has quit IRC | 08:50 | |
| *** baojg has quit IRC | 08:51 | |
| *** lpetrut has joined #openstack-manila | 09:18 | |
| *** rhagarty__ has joined #openstack-manila | 09:28 | |
| *** rhagarty_ has quit IRC | 09:31 | |
| *** baojg has joined #openstack-manila | 09:34 | |
| mkoderer__ | vponomaryov: regarding the hpb demo during midcycle | 09:40 |
| mkoderer__ | I added it on the etherpad | 09:40 |
| mkoderer__ | vponomaryov: but we can check how many folks are intressted since I already gave a short demo the netapp folks | 09:41 |
| mkoderer__ | I can also just do it for the ppl that are intressted in a separate meeting | 09:42 |
| vponomaryov | mkoderer__: wow, goog news | 09:46 |
| vponomaryov | mkoderer__: lots of people could just be shy to say it =) | 09:47 |
| mkoderer__ | vponomaryov: :) we will see | 09:53 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: Add dummy driver https://review.openstack.org/328391 | 09:56 |
| *** pcaruana has joined #openstack-manila | 10:09 | |
| *** baojg has quit IRC | 10:20 | |
| *** baojg has joined #openstack-manila | 10:32 | |
| *** baojg has quit IRC | 10:41 | |
| *** gouthamr has joined #openstack-manila | 10:43 | |
| vponomaryov | ganso: fake driver passed migration tests on your change - http://logs.openstack.org/91/328391/26/experimental/gate-manila-tempest-minimal-dsvm-dummy/310adf7/logs/tempest_2/testr_results.html.gz | 10:45 |
| vponomaryov | ganso: I guess, now you can finish tests for your change | 10:45 |
| *** yangyapeng has quit IRC | 10:51 | |
| ganso | vponomaryov: hi Valeriy | 10:55 |
| ganso | vponomaryov: I thought you said it was not good to have the driver returning the host | 10:56 |
| vponomaryov | ganso: yes | 10:56 |
| vponomaryov | ganso: you work on it? | 10:56 |
| ganso | vponomaryov: yes | 10:56 |
| vponomaryov | ok ) | 10:57 |
| ganso | vponomaryov: I am finishing the last steps | 10:59 |
| vponomaryov | goooood | 11:03 |
| vponomaryov | ganso: which migration driver methods called on source host and which on destination? | 11:32 |
| ganso | vponomaryov: only migration_get_driver_info is called on destination | 11:36 |
| ganso | vponomaryov: with the current patchset that is on gerrit I thought about migration_complete being able to be called on destination | 11:37 |
| ganso | vponomaryov: like, migration_complete would be called on whatever 'host' is in share_instance model | 11:37 |
| vponomaryov | ganso: either way it should be said explicitly in driver.py | 11:37 |
| vponomaryov | ganso: because logic of implementation changes | 11:38 |
| ganso | vponomaryov: so, if in migration_start driver updated host, migration_complete would be called on destination... this adds flexibility. I am removing this now, I don't know if it was necessary or not | 11:38 |
| ganso | vponomaryov: this was just introduced with this fix patch... that's why I have not updated yet | 11:38 |
| vponomaryov | ganso: also, the fact that only one share instance is used makes we do hacks | 11:38 |
| ganso | vponomaryov: yes, those hacks add flexibility, but they are hacks, that's why I am removing them | 11:39 |
| vponomaryov | ganso: ? | 11:39 |
| vponomaryov | ganso: I am saying driver should get two share instances - src and dest | 11:39 |
| ganso | vponomaryov: but I don't know if migration_complete ever needs to be invoked in the destination, maybe a vendor will say that it is needed and we will need to introduce this back | 11:39 |
| ganso | vponomaryov: yes, this is what I am implementing, thus, removing the hacks that drivers would return 'host' | 11:40 |
| vponomaryov | ganso: you are going to add it to existing "fix" commit? | 11:41 |
| ganso | vponomaryov: yes | 11:41 |
| vponomaryov | ganso: and you said you almost finished? | 11:41 |
| ganso | vponomaryov: not almost, I am in the last steps | 11:42 |
| ganso | vponomaryov: I still have to re-do unit tests and test thoroughly | 11:42 |
| ganso | vponomaryov: patch will not be completed today | 11:43 |
| ganso | vponomaryov: but I hope I can upload a testable version today, passing jenkins | 11:43 |
| vponomaryov | ganso: ok, then I am postponing work on ZFSonLinux implementation | 11:43 |
| vponomaryov | ganso: let me know if I can help you | 11:44 |
| ganso | vponomaryov: sure :) thanks | 11:44 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/python-manilaclient: Updated from global requirements https://review.openstack.org/332370 | 11:47 |
| *** timcl has joined #openstack-manila | 12:06 | |
| *** jseiler__ has joined #openstack-manila | 12:08 | |
| *** jseiler_ has quit IRC | 12:12 | |
| *** tpsilva has joined #openstack-manila | 12:19 | |
| openstackgerrit | Merged openstack/manila: Updated from global requirements https://review.openstack.org/331168 | 12:28 |
| *** cknight has joined #openstack-manila | 12:31 | |
| *** adrianofr has joined #openstack-manila | 12:38 | |
| *** cfey has quit IRC | 12:43 | |
| *** cknight has quit IRC | 12:45 | |
| *** erlon has joined #openstack-manila | 12:45 | |
| *** cfey has joined #openstack-manila | 12:48 | |
| *** cfey has quit IRC | 12:53 | |
| *** akerr has joined #openstack-manila | 12:54 | |
| *** cfey has joined #openstack-manila | 13:00 | |
| *** faiz89 has joined #openstack-manila | 13:04 | |
| bswartz | aovchinnikov: help -- ganesha is driving me crazy | 13:13 |
| aovchinnikov | bswartz: sure, what exactly is it doing? | 13:14 |
| *** akshai has joined #openstack-manila | 13:14 | |
| bswartz | every filesystem I mount is "Read only" and I can't see any files or write anything | 13:14 |
| bswartz | so something is wrong with the way I'm setting up access in ganesha.conf | 13:15 |
| aovchinnikov | do you run it on a real VM or in a container? | 13:15 |
| bswartz | I'm running Fedora 24 Server in a VM | 13:15 |
| bswartz | just trying to test ganesha in the most well supported environment | 13:15 |
| aovchinnikov | ok, so it is not a container privilege issue | 13:16 |
| aovchinnikov | could you please share your sample config? | 13:16 |
| aovchinnikov | also you could try disabling acl and switching off security just to test things | 13:16 |
| *** dustins has joined #openstack-manila | 13:18 | |
| aovchinnikov | that could be done by adding Disable_ACL = True; and SecType = "none"; fields to an export | 13:19 |
| *** merooney has joined #openstack-manila | 13:20 | |
| bswartz | aovchinnikov: http://paste.openstack.org/show/521935/ | 13:20 |
| bswartz | I've tried many variations, but the results are pretty much the same | 13:21 |
| aovchinnikov | ok, have you tried adding a client block? | 13:21 |
| *** porrua has joined #openstack-manila | 13:21 | |
| bswartz | when I attempt to change either "Path" or "Pseudo" to "/" though ganesha crashes are startup every time | 13:21 |
| bswartz | crashes at startup | 13:21 |
| openstackgerrit | Daniel Gonzalez Nothnagel proposed openstack/manila: Migrate API reference into tree https://review.openstack.org/313874 | 13:23 |
| aovchinnikov | like this: http://paste.openstack.org/show/Z4e3581Ib4VCmDrIUZW4/ | 13:23 |
| aovchinnikov | well, / is sort of illegal for both path and pseudo. but does not look like a good reason to crash | 13:24 |
| bswartz | many of the samples in the ganesha source code use / | 13:25 |
| bswartz | https://github.com/nfs-ganesha/nfs-ganesha/blob/master/src/config_samples/zfs.conf | 13:25 |
| bswartz | https://github.com/nfs-ganesha/nfs-ganesha/blob/master/src/config_samples/ceph.conf | 13:26 |
| *** akshai_ has joined #openstack-manila | 13:26 | |
| aovchinnikov | those are ceph and zfs, I am unsure about meaning of / in the context of ceph and zfs | 13:27 |
| bswartz | I shouldn't matter | 13:28 |
| aovchinnikov | I was referring to sharing local directories | 13:28 |
| bswartz | Path and Pseudo our outside the FSAL block | 13:28 |
| bswartz | are outside | 13:28 |
| aovchinnikov | yes, but shouldn't they be interpreted according to chosen abstraction layer? | 13:29 |
| bswartz | I dunno tbh | 13:29 |
| bswartz | the documentation is poor | 13:29 |
| aovchinnikov | because you'll end up sharing everything otherwise | 13:30 |
| *** akshai has quit IRC | 13:30 | |
| *** xyang1 has joined #openstack-manila | 13:30 | |
| aovchinnikov | which could be entertaining | 13:31 |
| *** dmk0202 has quit IRC | 13:31 | |
| *** cknight has joined #openstack-manila | 13:46 | |
| bswartz | aovchinnikov: my goal is not to share everything, but to allow clients to mount <ip>:/ and see all of things that are shared | 13:49 |
| aovchinnikov | yes, exactly | 13:50 |
| *** Yogi1 has joined #openstack-manila | 13:50 | |
| aovchinnikov | to do so you share a path to a folder you want to share, for example /shares/share1 | 13:50 |
| aovchinnikov | and a pseudopath | 13:50 |
| aovchinnikov | then you mount <ganesha_host_ip>:/ to a local directory | 13:51 |
| aovchinnikov | after which you can find pseudopath inside it | 13:51 |
| aovchinnikov | and all other shares allowed to the client's address | 13:52 |
| bswartz | aovchinnikov: I see you use quotes in your config file | 13:53 |
| bswartz | are those needed or optional? | 13:53 |
| aovchinnikov | docs say they are optional, I keep using them out of habit | 13:53 |
| *** shausy has quit IRC | 13:55 | |
| aovchinnikov | have tried out paths without quotes, everything works fine | 13:56 |
| bswartz | when I attempt to write a file to the share it just hangs | 13:57 |
| aovchinnikov | sometimes it takes several seconds to access a share | 13:57 |
| bswartz | does ganesha have a way to turn on verbose logging? | 13:57 |
| bswartz | the mount occurs somewhat quickly (like 2 seconds0 | 13:58 |
| aovchinnikov | it happens to me sporadically | 13:58 |
| aovchinnikov | yes | 13:58 |
| bswartz | oh it worked after like 15 seconds | 13:58 |
| bswartz | I wonder why it could be so slow | 13:58 |
| aovchinnikov | yes, it can take up to half a minute | 13:58 |
| bswartz | something is still strange | 13:58 |
| bswartz | what is the point of Disable_ACL = True; ? | 13:59 |
| aovchinnikov | -N NIV_DEBUG should set debug level | 14:00 |
| bswartz | where do you specify that | 14:00 |
| *** nkrinner is now known as nkrinner_afk | 14:00 | |
| aovchinnikov | it depends on how you start a daemon | 14:01 |
| *** MikeG451 has joined #openstack-manila | 14:01 | |
| bswartz | I'm just doing systemctl | 14:01 |
| aovchinnikov | ok, I start it manually: ganesha.nfsd -L /var/log/ganesha.log -F -N NIV_DEBUG | 14:03 |
| aovchinnikov | when I need to debug | 14:03 |
| aovchinnikov | it starts the daemon in foreground | 14:03 |
| aovchinnikov | writes extended log to specified log file | 14:03 |
| *** eharney has joined #openstack-manila | 14:07 | |
| *** yangyapeng has joined #openstack-manila | 14:08 | |
| *** yangyapeng has quit IRC | 14:08 | |
| aovchinnikov | I've set disable_acl to true to make things work without messing with nfsv4 acls. I was planning to remove it later from my export templates when basic ganesha helper works | 14:10 |
| *** timcl has quit IRC | 14:14 | |
| bswartz | what kind of failures did you see without that line? | 14:15 |
| bswartz | for some reason things are working a lot better now, albeit with weird I/O delays after I mount the FS | 14:16 |
| bswartz | also I found the place in the docs that describes using / as the export path | 14:17 |
| aovchinnikov | to be honest I don't remember. I am using it for quite a long time. I do remeber that you (probably) have to set up acl explicitly | 14:17 |
| bswartz | # Export_id (required)An identifier for the export, must be unique and | 14:18 |
| bswartz | #betweem 0 and 65535. If Export_Id 0 is specified, Pseudo | 14:18 |
| bswartz | # must be the root path (/). | 14:18 |
| bswartz | https://github.com/nfs-ganesha/nfs-ganesha/blob/master/src/config_samples/export.txt#L98 | 14:18 |
| aovchinnikov | also that will depend on a file system in use | 14:18 |
| aovchinnikov | that looks like a very special case IMO | 14:20 |
| bswartz | well it implies that when you don't define export ID 0, a special export is automatically used for / | 14:21 |
| bswartz | because NFSv4 requires that / be mountable | 14:21 |
| aovchinnikov | it looks like you can export something as / | 14:21 |
| aovchinnikov | but doesn't that mean that it will be the only thing exported? | 14:22 |
| bswartz | no, the magic of NFSv4 is that it contains a whole pseudo namespace which doesn't match the namespace of the host | 14:22 |
| aovchinnikov | because other possible pseudopathes should be placed somewhere inside pseudo-/ | 14:22 |
| bswartz | you can map parts of the hosts filesystem into differnet places in the pseudo filesystem | 14:22 |
| bswartz | NFSv4 exports can span filesystems too (very important) | 14:23 |
| aovchinnikov | yes, but as long as you export things as export 0 pseudo-/ represents the single physical exported directory, right? | 14:23 |
| bswartz | so I can create 2 xfs filesystems and mount them at /shares/one and /shares/two, but have a single NFS export at /shares | 14:23 |
| aovchinnikov | yes, but it still doesn't look like a good idea to mess with pseudo-/ | 14:23 |
| bswartz | clients who mount /shares will see both filesystems in the subdirectories | 14:24 |
| aovchinnikov | ok, what happens if someone defines pseudo to be "/shares/one"? | 14:24 |
| bswartz | then they can take something mounted elsewhere in the host's filesystem and make it looks like it's at /shares/one to the NFS client | 14:25 |
| aovchinnikov | so what happens to the original "/shares/one"? | 14:26 |
| bswartz | it's covered up and invisible to NFS clients | 14:26 |
| aovchinnikov | is it reacheable? is it shadowed? | 14:26 |
| bswartz | although the host (server) can still see it as usual | 14:26 |
| bswartz | yeah shadowed is the better technical term | 14:26 |
| aovchinnikov | looks like a great way to introduce some bugs and magic behavior to me | 14:27 |
| aovchinnikov | I've tried this one, now it shows me an empty export list, nevertheless I can mount something, which is not readable and not writeable | 14:35 |
| *** jcsp has joined #openstack-manila | 14:36 | |
| openstackgerrit | Daniel Gonzalez Nothnagel proposed openstack/manila: Migrate API reference into tree https://review.openstack.org/313874 | 14:36 |
| aovchinnikov | yet shadowing exports seem to work just fine | 14:37 |
| *** timcl has joined #openstack-manila | 14:43 | |
| bswartz | aovchinnikov: using showmount? | 14:43 |
| aovchinnikov | yes | 14:43 |
| bswartz | you should be aware that showmount is a NFSv3-only program | 14:44 |
| bswartz | it's completely unaware of NFSv4 filesystems because NFSv4 doesn't have a separate mount protocol like NFSv3 does | 14:44 |
| bswartz | with NFSv4 it's assumed that everyone can mount / and navigate down to the subdir they want to access | 14:44 |
| aovchinnikov | good point. | 14:45 |
| aovchinnikov | but I am not specifying the protocol to be used | 14:45 |
| aovchinnikov | if I get it right i t should produce a tag for v3 basing on path to shared directory | 14:46 |
| bswartz | well IMO we don't need to support v3 at all with this driver | 14:54 |
| bswartz | if we can get support for free I'm not opposed to it | 14:55 |
| bswartz | but if it's going to turn into a lot of work then I'd rather leave it off | 14:55 |
| bswartz | at least for the initial commit that adds ganesha | 14:55 |
| bswartz | I'm curious if ignoring v3 and focusing on v4 makes any of our ganesha-dbus interactions simpler or less problematic | 14:56 |
| aovchinnikov | bswartz: dbus interactions are protocol-independent | 15:07 |
| aovchinnikov | and there are not so many options for talking to it via dbus anyway | 15:08 |
| aovchinnikov | we still have to create export configs inside a container, and adding it is pretty much independent of what is inside the config | 15:09 |
| *** eharney has quit IRC | 15:12 | |
| *** faiz89 has quit IRC | 15:14 | |
| *** baojg has joined #openstack-manila | 15:19 | |
| *** eharney has joined #openstack-manila | 15:27 | |
| openstackgerrit | Marcus V R Nascimento proposed openstack/manila: Add share_size config option https://review.openstack.org/333967 | 15:29 |
| *** pcaruana has quit IRC | 15:38 | |
| *** gouthamr has quit IRC | 15:39 | |
| *** timcl has quit IRC | 15:39 | |
| *** gouthamr has joined #openstack-manila | 15:39 | |
| *** jcsp has quit IRC | 15:44 | |
| openstackgerrit | Merged openstack/python-manilaclient: [CI] Fix devstack hooks to unblock CI https://review.openstack.org/333229 | 15:46 |
| *** sgotliv__ has joined #openstack-manila | 15:52 | |
| *** Suyash has joined #openstack-manila | 15:54 | |
| *** sgotliv__ has quit IRC | 16:01 | |
| vponomaryov | bswartz: manila mitaka branch fix is waiting for your attention - https://review.openstack.org/#/c/332883/ | 16:03 |
| *** akshai_ has quit IRC | 16:11 | |
| *** timcl has joined #openstack-manila | 16:21 | |
| *** gouthamr has quit IRC | 16:23 | |
| *** eharney has quit IRC | 16:25 | |
| *** baojg has quit IRC | 16:26 | |
| *** akshai has joined #openstack-manila | 16:28 | |
| *** dmk0202 has joined #openstack-manila | 16:34 | |
| *** eharney has joined #openstack-manila | 16:38 | |
| *** gouthamr has joined #openstack-manila | 16:44 | |
| gouthamr | xyang1 markstur bswartz toabctl : would appreciate a +2/+W on https://review.openstack.org/#/c/323646/ :) | 16:48 |
| *** pgbridge has joined #openstack-manila | 16:52 | |
| xyang1 | gouthamr: I'll take a look | 16:52 |
| gouthamr | xyang1: thank you! | 16:53 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila: [ZFSonLinux] Add 'manage share' feature support https://review.openstack.org/333994 | 16:54 |
| openstackgerrit | Merged openstack/manila-specs: Extend the design of share networks to span subnets https://review.openstack.org/323646 | 17:00 |
| *** akshai has quit IRC | 17:02 | |
| *** akshai has joined #openstack-manila | 17:02 | |
| *** faiz89 has joined #openstack-manila | 17:07 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila-ui: Fix error handling for 'extend' share operation https://review.openstack.org/319377 | 17:09 |
| *** dmk0202 has quit IRC | 17:15 | |
| *** catintheroof has joined #openstack-manila | 17:16 | |
| *** dmk0202 has joined #openstack-manila | 17:16 | |
| *** rraja has quit IRC | 17:16 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 17:18 |
| openstackgerrit | Alexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor https://review.openstack.org/309286 | 17:23 |
| *** cbader has quit IRC | 17:24 | |
| *** alkhodos has joined #openstack-manila | 17:29 | |
| openstackgerrit | Ramana Raja proposed openstack/manila: cephfs_native: add read-only support https://review.openstack.org/334010 | 17:30 |
| alkhodos | Hi everyone. My CI hit this issue http://paste.openstack.org/show/522001/ with credentials and I cannot really understand which credentials are we talking about here. Any suggestions what's wrong or probably misconfigured? | 17:32 |
| *** dmk0202 has quit IRC | 17:35 | |
| openstackgerrit | Merged openstack/python-manilaclient: manila list --all-tenants should display projectID https://review.openstack.org/312017 | 17:39 |
| *** dmk0202 has joined #openstack-manila | 17:39 | |
| *** cbader has joined #openstack-manila | 17:40 | |
| *** lpetrut has quit IRC | 17:46 | |
| *** akshai has quit IRC | 17:48 | |
| *** akshai has joined #openstack-manila | 17:49 | |
| tpsilva | xyang1: ping | 17:51 |
| tpsilva | anyone familiar with the scheduler/thin provisioning? | 17:54 |
| tpsilva | https://github.com/openstack/manila/blob/master/manila/scheduler/host_manager.py#L330 | 17:54 |
| tpsilva | is this is wrong? | 17:54 |
| *** DaiDV has joined #openstack-manila | 17:58 | |
| *** DaiDV has quit IRC | 17:58 | |
| *** openstackgerrit has quit IRC | 18:03 | |
| *** openstackgerrit has joined #openstack-manila | 18:03 | |
| xyang1 | tpsilva: not wrong | 18:05 |
| tpsilva | xyang1: but if a backend is thin provisioned, this is reserving the whole share size | 18:05 |
| xyang1 | tpsilva: I actually has a discussion with winston-d on this a while back. he can probably explain better. | 18:07 |
| xyang1 | tpsilva: should have added some notes about this | 18:07 |
| xyang1 | tpsilva: this is to prevent overprovisioning, it will be updated all the time | 18:08 |
| tpsilva | xyang1: but when the driver update_stats is called, this value is updated with the real backend stats, right? | 18:08 |
| xyang1 | tpsilva: I'll try to find the patch where someone was trying to remove this, and there were comments fro winston-d to explain wht it is there | 18:09 |
| gouthamr | xyang1: you mean this one? https://review.openstack.org/#/c/271172/7 | 18:10 |
| xyang1 | gouthamr: no | 18:10 |
| gouthamr | probably not, no comments there.. | 18:10 |
| xyang1 | gouthamr: there was a patch long time ago in cinder | 18:11 |
| gouthamr | xyang1: ah.. okay, nvm :) | 18:11 |
| gouthamr | xyang1: probably this one: https://review.openstack.org/#/c/189261/ | 18:12 |
| tpsilva | I was looking through the abandoned patches, but cinder has too much | 18:12 |
| tpsilva | xyang1: cool, let me check | 18:12 |
| tpsilva | it looks really wrong, but cinder has this exact same line | 18:15 |
| *** eharney has quit IRC | 18:15 | |
| xyang1 | tpsilva: winston-d explained to me why it was there but I don't remember the exact reason | 18:15 |
| tpsilva | xyang1: is this discussion logged anywhere? | 18:16 |
| xyang1 | tpsilva: I don't know. too long ago | 18:17 |
| xyang1 | tpsilva: I'll see if I can find it. if not, I'll ping him | 18:17 |
| tpsilva | xyang1: I'm curious because this is preventing the CI for our new driver to run successfully... it tries to create several shares and eventually the free space becomes negative, since we need to create 128GB shares | 18:17 |
| tpsilva | xyang1: we can simply expand our pool, but that doesn't look right | 18:18 |
| tpsilva | xyang1: but thanks! I'll check with winston-d eventually | 18:18 |
| xyang1 | tpsilva: this could be to prevent overprovisioning | 18:19 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/python-manilaclient: Data Service CLI commands https://review.openstack.org/333329 | 18:20 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/python-manilaclient: Update Share Migration CLI https://review.openstack.org/328437 | 18:20 |
| xyang1 | tpsilva: there is no way to know the exact capacity all the time | 18:20 |
| xyang1 | tpsilva: I think this is only effective for a brief period before the driver send update stats again | 18:21 |
| tpsilva | xyang1: it makes sense... it avoids several shares/volumes from being created on the same backend all at once | 18:23 |
| tpsilva | xyang1: I was talking to erlon, he said that we can change the update_stats timeout... do you know if manila implements this parameter? | 18:24 |
| xyang1 | tpsilva: this line of code was added on 10/2, 2012:) | 18:24 |
| xyang1 | tpsilva: when winston-d first added filter scheduler in cinder | 18:25 |
| xyang1 | tpsilva: so this logic was probably from nova | 18:25 |
| tpsilva | xyang1: yep, it makes sense | 18:26 |
| tpsilva | xyang1: do you know what parameter is that? | 18:26 |
| xyang1 | tpsilva: which one? | 18:26 |
| tpsilva | xyang1: I'm looking on cinder... don't know if manila implements that | 18:26 |
| tpsilva | xyang1: to change the update_stats timeout | 18:26 |
| xyang1 | tpsilva: It may be ported to manila but I need to check | 18:27 |
| tpsilva | xyang1: periodic_interval | 18:29 |
| tpsilva | xyang1: found it | 18:29 |
| xyang1 | ok | 18:29 |
| tpsilva | xyang1: thanks a lot! | 18:29 |
| xyang1 | tpsilva: welcome | 18:29 |
| *** eharney has joined #openstack-manila | 18:34 | |
| *** openstackstatus has quit IRC | 18:40 | |
| *** openstack has joined #openstack-manila | 18:44 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: PoC Data Service API https://review.openstack.org/333020 | 18:46 |
| *** gouthamr has quit IRC | 18:51 | |
| *** lpetrut has joined #openstack-manila | 18:55 | |
| *** gouthamr has joined #openstack-manila | 19:00 | |
| openstackgerrit | Merged openstack/manila: Update quota usages correctly in manage share operation https://review.openstack.org/333199 | 19:07 |
| *** faiz89 has quit IRC | 19:31 | |
| *** dustins has joined #openstack-manila | 19:33 | |
| openstackgerrit | Merged openstack/manila: [Tempest] Remove noqa filters https://review.openstack.org/329355 | 19:44 |
| *** lpetrut has quit IRC | 19:50 | |
| openstackgerrit | Diem Tran proposed openstack/manila: Add share manage/unmanage of Oracle ZFSSA driver https://review.openstack.org/327897 | 19:50 |
| *** merooney has quit IRC | 19:52 | |
| openstackgerrit | Diem Tran proposed openstack/manila: Add share manage/unmanage of Oracle ZFSSA driver https://review.openstack.org/327897 | 19:52 |
| *** Yogi1 has quit IRC | 19:52 | |
| openstackgerrit | Diem Tran proposed openstack/manila: Add share manage/unmanage of Oracle ZFSSA driver https://review.openstack.org/327897 | 19:55 |
| bswartz | tbarron: ping | 19:56 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 19:57 |
| tbarron | bswartz: you shoud be at beer bash | 19:57 |
| bswartz | tbarron: indeed, but I'm working from home today .... a poor choice of days to work for home | 19:57 |
| * tbarron fondly remembers a little shared scotch | 19:58 | |
| tbarron | bswartz: what's up? | 19:58 |
| *** akerr has quit IRC | 20:03 | |
| *** porrua has quit IRC | 20:21 | |
| *** akshai has quit IRC | 20:23 | |
| *** Suyash has quit IRC | 20:31 | |
| *** Suyash has joined #openstack-manila | 20:32 | |
| *** Suyash has quit IRC | 20:32 | |
| *** Suyash has joined #openstack-manila | 20:32 | |
| *** timcl has quit IRC | 20:35 | |
| *** faiz89 has joined #openstack-manila | 20:45 | |
| *** dmk0202 has quit IRC | 20:46 | |
| *** faiz89 has quit IRC | 21:19 | |
| *** gouthamr has quit IRC | 21:36 | |
| *** dustins has quit IRC | 21:52 | |
| *** gouthamr has joined #openstack-manila | 21:53 | |
| *** gouthamr_ has joined #openstack-manila | 21:55 | |
| *** gouthamr has quit IRC | 21:58 | |
| *** adrianofr has quit IRC | 22:01 | |
| *** alyson_ has quit IRC | 22:03 | |
| *** harlowja has quit IRC | 22:17 | |
| *** tpsilva has quit IRC | 22:18 | |
| *** catintheroof has quit IRC | 22:18 | |
| *** harlowja has joined #openstack-manila | 22:30 | |
| *** MikeG451 has quit IRC | 22:56 | |
| *** pgbridge has quit IRC | 23:20 | |
| *** cknight has quit IRC | 23:28 | |
| *** faiz89 has joined #openstack-manila | 23:52 | |
| *** faiz891 has joined #openstack-manila | 23:55 | |
| *** faiz89 has quit IRC | 23:57 | |
| *** faiz891 has quit IRC | 23:59 | |
| *** faiz89 has joined #openstack-manila | 23:59 | |
| *** faiz89 has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!