| *** catintheroof has joined #openstack-manila | 00:04 | |
| *** porrua has quit IRC | 00:08 | |
| *** tinwood has quit IRC | 00:10 | |
| *** tinwood has joined #openstack-manila | 00:11 | |
| *** porrua has joined #openstack-manila | 00:19 | |
| bswartz | ganso: https://bugs.launchpad.net/manila/+bug/1661381 | 00:25 |
|---|---|---|
| openstack | Launchpad bug 1661381 in Manila "Can't update snapshot instance export locations on completion of (driver assisted) migration" [High,New] | 00:25 |
| ganso | bswartz: yes? | 00:25 |
| bswartz | gouthamr ganso: is one of you working on this? | 00:25 |
| ganso | bswartz: I am working on it right now | 00:26 |
| bswartz | please assign yourself | 00:26 |
| ganso | bswartz: sure, sorry | 00:26 |
| bswartz | I'll leave it targeted because it's actively being worked on | 00:26 |
| *** porrua has quit IRC | 00:26 | |
| bswartz | gouthamr: https://bugs.launchpad.net/manila/+bug/1661421 | 00:26 |
| openstack | Launchpad bug 1661421 in Manila "Sporadic share type cleanup failure in tests in ShareIpRulesForNFSNegativeTest" [High,New] - Assigned to Goutham Pacha Ravi (gouthamr) | 00:26 |
| bswartz | gouthamr: ^ this bug has no proposed fix, are you working on one now? | 00:27 |
| bswartz | ganso: the LVM revert fix passed my testing: https://review.openstack.org/428398 | 00:27 |
| bswartz | ganso: at this time it does seem to cause the mount to go stale | 00:28 |
| bswartz | ganso: other than that the behaviour appears correct | 00:29 |
| ganso | bswartz: that's much better than getting stuck in lv merge | 00:29 |
| ganso | bswartz: for HNAS driver, the same happens if user is mounted while it is reverting... I guess this is inevitable | 00:29 |
| ganso | bswartz: can NetApp driver revert with user connected and not having to remount? | 00:30 |
| bswartz | ganso: I think it ultimately depends on whether the revert operation changes the inode of the root | 00:30 |
| *** catintheroof has quit IRC | 00:30 | |
| bswartz | ganso yes | 00:30 |
| ganso | bswartz: whoa cool | 00:30 |
| bswartz | it's simple magic in nfsd | 00:30 |
| *** porrua has joined #openstack-manila | 00:30 | |
| openstackgerrit | Merged openstack/manila master: Fix HNAS driver always handling mountable snapshots https://review.openstack.org/428077 | 00:32 |
| ganso | markstur: we can +W another one ^ | 00:33 |
| markstur | ganso: which? | 00:33 |
| ganso | markstur: gerrit hasn't reported conflicts yet, let me try cherry picks here and see which one does not conflict | 00:34 |
| markstur | the one at ^ is merged. I'll look at the others | 00:34 |
| *** sapcc-bot has quit IRC | 00:34 | |
| *** sapcc-bot has joined #openstack-manila | 00:35 | |
| ganso | markstur: https://review.openstack.org/#/c/428267 | 00:35 |
| ganso | markstur: yea, that's what I meant | 00:35 |
| *** porrua has quit IRC | 00:35 | |
| ganso | markstur: that one does not conflict ^ | 00:35 |
| ganso | markstur: ok I was able to cherry pick all of them, no conflicts | 00:36 |
| markstur | the 3rdparty CIs are about 50/50 pass/fail on that one | 00:37 |
| ganso | markstur: well, the change doesn't impact them... I guess several 3rd party CIs are not working | 00:41 |
| markstur | ganso: I was thinking something like since yours works now, you should have to fix one other one each time you need to land a patch | 00:42 |
| markstur | ... but that seems a bit harsh. You should be rewarded. | 00:43 |
| ganso | lol | 00:43 |
| markstur | ganso: https://review.openstack.org/#/c/428398/ | 00:45 |
| ganso | markstur: yes? | 00:45 |
| markstur | ganso: I suppose it is after hours for you, but that ^ could use review. | 00:46 |
| ganso | markstur: yea I am working on https://bugs.launchpad.net/manila/+bug/1661381 right now, will submit patch soon then I'll review that one | 00:46 |
| openstack | Launchpad bug 1661381 in Manila "Can't update snapshot instance export locations on completion of (driver assisted) migration" [High,New] - Assigned to Rodrigo Barbieri (rodrigo-barbieri2010) | 00:46 |
| *** sapcc-bot has quit IRC | 00:54 | |
| *** sapcc-bot has joined #openstack-manila | 00:54 | |
| openstackgerrit | Yogesh proposed openstack/manila master: Improve test coverage for share migration https://review.openstack.org/418559 | 00:57 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: [WIP] Fix migration of mountable snapshots https://review.openstack.org/428526 | 01:08 |
| ganso | markstur: reviewed | 01:19 |
| *** crushil has joined #openstack-manila | 01:20 | |
| ganso | markstur: I'd review vponomaryov's change, but I had already taken a look at it and judging by the current comments, it still looks a bit confusing wrt to what it is supposed to be including in the patch | 01:21 |
| markstur | yep | 01:26 |
| *** xyang_ has joined #openstack-manila | 01:31 | |
| *** gcb has joined #openstack-manila | 01:31 | |
| *** xyang_ has quit IRC | 01:32 | |
| *** kberger has quit IRC | 01:44 | |
| *** gouthamr has joined #openstack-manila | 02:08 | |
| openstackgerrit | Tom Barron proposed openstack/manila master: Do not merge - baseline https://review.openstack.org/428463 | 02:09 |
| openstackgerrit | Tom Barron proposed openstack/manila master: Do Not Merge - new tempest https://review.openstack.org/428230 | 02:10 |
| *** xinyanzhang has quit IRC | 02:16 | |
| *** xinyanzhang has joined #openstack-manila | 02:22 | |
| *** mtanino has quit IRC | 02:28 | |
| *** xyang_ has joined #openstack-manila | 02:46 | |
| *** crushil has quit IRC | 03:05 | |
| *** crushil has joined #openstack-manila | 03:25 | |
| *** xyang_ has quit IRC | 03:33 | |
| *** xyang_ has joined #openstack-manila | 03:34 | |
| *** xinyanzhang has quit IRC | 03:35 | |
| *** eharney has quit IRC | 03:40 | |
| *** crushil has quit IRC | 03:41 | |
| *** ganso has quit IRC | 03:44 | |
| *** crushil has joined #openstack-manila | 04:07 | |
| *** gouthamr has quit IRC | 04:17 | |
| *** furlongm has quit IRC | 04:27 | |
| *** sticker__ has joined #openstack-manila | 04:29 | |
| *** furlongm has joined #openstack-manila | 04:32 | |
| *** sticker_ has quit IRC | 04:32 | |
| *** crushil has quit IRC | 04:33 | |
| *** crushil has joined #openstack-manila | 04:33 | |
| *** crushil has quit IRC | 04:39 | |
| *** crushil has joined #openstack-manila | 04:41 | |
| openstackgerrit | zhongjun proposed openstack/manila master: Enable IPv6 in manila(allow access) https://review.openstack.org/312321 | 04:53 |
| *** akapil has joined #openstack-manila | 05:08 | |
| *** shausy has joined #openstack-manila | 05:09 | |
| *** akapil has quit IRC | 05:13 | |
| *** crushil has quit IRC | 05:16 | |
| *** crushil has joined #openstack-manila | 05:16 | |
| *** shausy has quit IRC | 05:25 | |
| *** shausy has joined #openstack-manila | 05:26 | |
| *** lpetrut has joined #openstack-manila | 05:32 | |
| *** sticker__ has quit IRC | 05:32 | |
| *** xyang_ has quit IRC | 05:41 | |
| *** crushil has quit IRC | 06:13 | |
| *** lpetrut has quit IRC | 06:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila-ui master: Imported Translations from Zanata https://review.openstack.org/428607 | 07:18 |
| *** dsariel has joined #openstack-manila | 07:23 | |
| *** nkrinner_afk is now known as nkrinner | 07:23 | |
| *** xinyanzhang has joined #openstack-manila | 07:33 | |
| *** kaisers2 has quit IRC | 07:34 | |
| *** kaisers1 has quit IRC | 07:34 | |
| *** kaisers has joined #openstack-manila | 07:38 | |
| *** lpetrut has joined #openstack-manila | 07:45 | |
| *** FL1SK has quit IRC | 08:32 | |
| *** wlhc has joined #openstack-manila | 08:36 | |
| *** wlhc has quit IRC | 08:37 | |
| openstackgerrit | TommyLike proposed openstack/manila-specs master: Add spec for share backup https://review.openstack.org/330306 | 08:46 |
| openstackgerrit | TommyLike proposed openstack/manila-specs master: Add spec for share backup https://review.openstack.org/330306 | 08:49 |
| *** akapil has joined #openstack-manila | 09:04 | |
| *** akapil has quit IRC | 09:08 | |
| *** akapil has joined #openstack-manila | 09:08 | |
| tommylikehu | hey vponomaryov are you around? | 09:33 |
| vponomaryov | tommylikehu: I am instead of him today | 09:34 |
| *** a-pugachev has joined #openstack-manila | 09:36 | |
| tommylikehu | vponomaryov: oh, what do you mean by instead of him, are you not vponomaryov? | 09:36 |
| vponomaryov | vponomaryov: ^_^ | 09:37 |
| vponomaryov | tommylikehu: just kidding | 09:37 |
| vponomaryov | tommylikehu: what did you want to ask? | 09:37 |
| *** chlong has quit IRC | 09:38 | |
| tommylikehu | vponomaryov: about the comment here: https://review.openstack.org/#/c/406776/28/manila/share/drivers/netapp/dataontap/cluster_mode/drv_single_svm.py, even I move the method to the base driver, a lot of existing driver still needed to updated to report IP version capabilities, and we should guarantee that base.add_ip_version_capability is called after that, am I correct? | 09:42 |
| tommylikehu | vponomaryov: please forward this message to him, :) | 09:42 |
| *** rraja has joined #openstack-manila | 09:43 | |
| vponomaryov | tommylikehu: just show your code | 09:43 |
| vponomaryov | tommylikehu: I do not see any problems in setting default capability | 09:44 |
| vponomaryov | for all drivers | 09:44 |
| tommylikehu | even for driver CephFS Native? | 09:45 |
| vponomaryov | tommylikehu: црфе шы цкщтп цшер ше, | 09:47 |
| vponomaryov | tommylikehu: what iswrong with it? | 09:47 |
| vponomaryov | tommylikehu: every driver uses same approach for reporting capabilities | 09:48 |
| tommylikehu | vponomaryov: or we can have the default true for IPv4 and false for IPv6 and then modify these non-IP drivers set IPv4 and IPv6 both false? | 09:48 |
| vponomaryov | tommylikehu: is your spec about "access rules" or possibility to have networking of that type? | 09:49 |
| vponomaryov | tommylikehu: looks like your extra specs now sound ambiguous | 09:50 |
| tommylikehu | these two extra_specs only works when the access type is IP | 09:51 |
| tommylikehu | so the capabilities only represent the capbilities about access type IP | 09:52 |
| *** ganso has joined #openstack-manila | 09:52 | |
| vponomaryov | tommylikehu: it is not clear from name of specs | 09:54 |
| vponomaryov | tommylikehu: what if I want to have that type of networking? | 09:54 |
| vponomaryov | tommylikehu: nut don't care about type of rules | 09:54 |
| vponomaryov | s/nut/but/ | 09:54 |
| vponomaryov | цшдд иу ифсл дфеук | 09:56 |
| vponomaryov | will be back later | 09:56 |
| *** erlon has joined #openstack-manila | 09:57 | |
| tommylikehu | vponomaryov:ok | 09:57 |
| *** yumiriam has joined #openstack-manila | 09:57 | |
| tommylikehu | with that, this feature could be more complicated and more details are involoved, especially for the access_rule patches | 09:59 |
| tommylikehu | vponomaryov: I think the spec: https://github.com/openstack/manila-specs/blob/master/specs/ocata/manila-ipv6.rst, already mentioned that we gonna use this capabilities for IP based drivers. We also have implementations based on this presets at out patches. | 10:01 |
| tommylikehu | vponomaryov: hope bswartz and tbarron have noticed this. | 10:02 |
| openstackgerrit | Merged openstack/manila-ui master: Imported Translations from Zanata https://review.openstack.org/428607 | 10:03 |
| *** igajsin1 has joined #openstack-manila | 10:04 | |
| *** igajsin1 has left #openstack-manila | 10:11 | |
| *** carthaca_ has joined #openstack-manila | 10:21 | |
| *** akapil has quit IRC | 10:26 | |
| *** akapil has joined #openstack-manila | 10:28 | |
| *** FL1SK has joined #openstack-manila | 10:36 | |
| ganso | vponomaryov: ping | 11:00 |
| *** alyson_ has joined #openstack-manila | 11:04 | |
| *** akapil has quit IRC | 11:22 | |
| *** jprovazn has joined #openstack-manila | 11:25 | |
| *** akapil has joined #openstack-manila | 11:43 | |
| *** akapil has quit IRC | 11:47 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix migration of mountable snapshots https://review.openstack.org/428526 | 11:47 |
| *** akapil has joined #openstack-manila | 11:50 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix nonsense variable name https://review.openstack.org/428707 | 11:56 |
| *** crushil has joined #openstack-manila | 11:56 | |
| *** akapil_ has joined #openstack-manila | 12:02 | |
| *** eharney has joined #openstack-manila | 12:04 | |
| *** akapil has quit IRC | 12:04 | |
| *** akapil_ has quit IRC | 12:05 | |
| *** akapil has joined #openstack-manila | 12:15 | |
| *** gcb has quit IRC | 12:19 | |
| *** catintheroof has joined #openstack-manila | 12:20 | |
| *** catintheroof has quit IRC | 12:20 | |
| *** catintheroof has joined #openstack-manila | 12:21 | |
| *** akapil has joined #openstack-manila | 12:34 | |
| *** akapil has quit IRC | 12:36 | |
| *** mkoderer has joined #openstack-manila | 12:49 | |
| *** akapil has joined #openstack-manila | 12:52 | |
| vponomaryov | ganso: pong | 13:10 |
| *** gaurangt has quit IRC | 13:11 | |
| *** gaurangt has joined #openstack-manila | 13:11 | |
| vponomaryov | tommylikehu: it does not matter what was merged in specs if something is confusing for everyone | 13:11 |
| vponomaryov | tommylikehu: I mentioned use case, I am user, I want share in IPv6 network, but don't care about type of access rules, what should I do? | 13:12 |
| vponomaryov | tommylikehu: i see capability "ipv6_support=True" and think it satisfies my criterion | 13:13 |
| vponomaryov | tommylikehu: but it is not | 13:13 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Fix default approach for share group snapshot creation https://review.openstack.org/427811 | 13:31 |
| *** crushil has quit IRC | 13:32 | |
| vponomaryov | markstur, gouthamr: in https://review.openstack.org/#/c/427811 you request for unrelated to this "fix" things | 13:32 |
| vponomaryov | markstur, gouthamr: such as SG capability of consistent snapshots and update of API | 13:32 |
| vponomaryov | gouthamr: "addon of capability of consistent snapshots to SG model", IMHO, is update of code for ease of future backports, should be done separately. | 13:34 |
| *** hoonetorg has quit IRC | 13:35 | |
| vponomaryov | markstur: same about API update - "provider_location" is required for drivers to "delete" resources on "delete" call. View of this data in API is not required to fix this bug. | 13:35 |
| *** xyang1 has joined #openstack-manila | 13:38 | |
| *** hoonetorg has joined #openstack-manila | 13:48 | |
| *** crushil has joined #openstack-manila | 13:53 | |
| *** kberger has joined #openstack-manila | 13:56 | |
| *** crushil has quit IRC | 14:00 | |
| openstackgerrit | Yogesh proposed openstack/manila master: Improve test coverage for share migration https://review.openstack.org/418559 | 14:00 |
| *** ociuhandu has quit IRC | 14:01 | |
| *** Yogi1 has joined #openstack-manila | 14:04 | |
| *** crushil has joined #openstack-manila | 14:12 | |
| *** gouthamr has joined #openstack-manila | 14:14 | |
| *** eharney has quit IRC | 14:15 | |
| *** xyang_ has joined #openstack-manila | 14:21 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Add 'consistent_snapshot_support' attr to 'share_groups' DB model https://review.openstack.org/428748 | 14:21 |
| *** dustins has joined #openstack-manila | 14:24 | |
| ganso | vponomaryov: Hi Valeriy | 14:26 |
| ganso | vponomaryov: could you please take a look at https://review.openstack.org/#/c/427660/ and https://review.openstack.org/#/c/428267 ? they are ready | 14:26 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Add 'consistent_snapshot_support' attr to 'share_groups' DB model https://review.openstack.org/428748 | 14:28 |
| ganso | Yogi1, gouthamr: ping | 14:32 |
| gouthamr | ganso: pong | 14:32 |
| *** eharney has joined #openstack-manila | 14:32 | |
| ganso | gouthamr: I tested https://review.openstack.org/#/c/428526 and it fixed the problem, maybe you can rebase on top of it | 14:32 |
| Yogi1 | ganso Hey | 14:33 |
| ganso | Yogi1: ^ | 14:33 |
| Yogi1 | ganso ok sure | 14:33 |
| *** akapil has quit IRC | 14:36 | |
| *** shausy has quit IRC | 14:39 | |
| *** eharney has quit IRC | 14:40 | |
| gouthamr | ganso: looks good. some nits. we'll do a depends-on | 14:41 |
| *** gaurangt_ has joined #openstack-manila | 14:42 | |
| ganso | gouthamr: replied to your comment, do you suggest any change? so I can include in the new patch | 14:43 |
| gouthamr | ganso: can't see your reply.. | 14:44 |
| *** gaurangt has quit IRC | 14:44 | |
| *** kaisers has quit IRC | 14:45 | |
| *** alyson_ has quit IRC | 14:47 | |
| *** akapil has joined #openstack-manila | 14:52 | |
| *** mtanino has joined #openstack-manila | 14:52 | |
| ganso | gouthamr: yes, because I did not publish it xD | 14:53 |
| ganso | gouthamr: sorry | 14:53 |
| ganso | gouthamr: it is up now | 14:53 |
| *** eharney has joined #openstack-manila | 14:54 | |
| *** mtanino_ has joined #openstack-manila | 15:01 | |
| *** mtanino has quit IRC | 15:02 | |
| *** eharney has quit IRC | 15:02 | |
| openstackgerrit | Goutham Pacha Ravi proposed openstack/manila master: Fix wrong access-rule negative test https://review.openstack.org/428771 | 15:03 |
| *** yumiriam has quit IRC | 15:08 | |
| *** yumiriam has joined #openstack-manila | 15:09 | |
| *** mtanino_ has quit IRC | 15:09 | |
| *** pcaruana has quit IRC | 15:11 | |
| *** mtanino has joined #openstack-manila | 15:11 | |
| *** pcaruana has joined #openstack-manila | 15:17 | |
| *** eharney has joined #openstack-manila | 15:18 | |
| *** nkrinner is now known as nkrinner_afk | 15:23 | |
| *** chlong has joined #openstack-manila | 15:23 | |
| bswartz | tommylikehu: ping | 15:29 |
| bswartz | tommylikehu: you're probably not still around this late, but I added more notes in the Pike PTG etherpad about ipv6 | 15:30 |
| bswartz | #link https://etherpad.openstack.org/p/manila-pike-ptg-topics | 15:30 |
| bswartz | we've had multiple discussions about the access rules and capabilities but I don't think we ever settled on a final solution | 15:31 |
| bswartz | ganso: ping | 15:34 |
| ganso | bswartz: in a meeting | 15:36 |
| ganso | bswartz: for the next 3 hours | 15:36 |
| bswartz | ganso: >_< | 15:37 |
| ganso | bswartz: I'll respond whenever I can | 15:37 |
| ganso | bswartz: what's it? | 15:37 |
| bswartz | ganso: https://review.openstack.org/#/c/428398/4/manila/share/manager.py | 15:37 |
| tbarron | bswartz: tempest 15.0 will be getting cut sometime relatively soon & I'm looking at re-pinning manila from 12.0 | 15:43 |
| tbarron | bswartz: atm I have a review with current tempest commit and a baseline review | 15:44 |
| tbarron | bswartz: the current one passes, but there's a lot of variation in failures in the non-voting jobs in both the baseline and the current pin versions | 15:44 |
| tbarron | bswartz: so I will study output and look for gotchas | 15:45 |
| tbarron | bswartz: how long will we have from a manila standpoint before final decison on whether we can re-pin and to what point? | 15:45 |
| tbarron | for reference: | 15:46 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Add 'consistent_snapshot_support' attr to 'share_groups' DB model https://review.openstack.org/428748 | 15:46 |
| tbarron | https://review.openstack.org/#/c/428463 | 15:46 |
| tbarron | https://review.openstack.org/#/c/428230 | 15:46 |
| tbarron | hmm, bot seems busted | 15:47 |
| tbarron | well, gone | 15:47 |
| tbarron | vponomaryov: see questions about tempest repin above when you have time. | 15:48 |
| *** akapil has quit IRC | 15:54 | |
| *** crushil has quit IRC | 16:06 | |
| *** crushil has joined #openstack-manila | 16:06 | |
| *** rraja has quit IRC | 16:25 | |
| ganso | bswartz: I've taken a quick look and still think it would be better to use the share instance | 16:35 |
| *** chlong has quit IRC | 16:36 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix migration of mountable snapshots https://review.openstack.org/428526 | 16:38 |
| ganso | bswartz: I've suggested how to do it in the comment, we have access to the share instance | 16:39 |
| ganso | gouthamr: Could you please take a look at https://review.openstack.org/#/c/427660/ and https://review.openstack.org/#/c/428267/ ? They are ready to be workflowed I believe, they were workflowed yesterday but had to fix conflicts during merge | 16:41 |
| ganso | vponomaryov: thanks | 16:49 |
| vponomaryov | ganso: I am not gouthamr , but I hope you are not upset )) | 16:49 |
| ganso | vponomaryov: definitely not, I am very happy, thanks! | 16:50 |
| gouthamr | vponomaryov: thanks :P | 16:50 |
| *** lpetrut has quit IRC | 16:50 | |
| gouthamr | ganso vponomaryov: some busy day today.. | 16:50 |
| vponomaryov | tbarron: how can I help you? | 16:50 |
| vponomaryov | tbarron: I am for pinning tempest version in O | 16:51 |
| vponomaryov | we use too old version now | 16:51 |
| vponomaryov | 5 month old | 16:51 |
| tbarron | vponomaryov: I'm not trying to remove the pin, just to update it | 16:51 |
| vponomaryov | tbarron: I understand | 16:51 |
| tbarron | vponomaryov: and just wanted to make sure it was done in your view :D | 16:51 |
| *** chlong has joined #openstack-manila | 16:52 | |
| vponomaryov | tbarron: do oyu need help with it? | 16:52 |
| tbarron | vponomaryov: no, it was just and FYI/heads up | 16:52 |
| vponomaryov | tbarron: ok | 16:53 |
| *** mtanino has quit IRC | 16:53 | |
| *** mtanino has joined #openstack-manila | 16:53 | |
| tbarron | vponomaryov: so far updating the pin looks pretty good, but I want to do it with the actual commit that tempest releases at if possible | 16:53 |
| *** mtanino has quit IRC | 16:53 | |
| tbarron | vponomaryov: hence my questions to bswartz about how late we can do this | 16:53 |
| vponomaryov | markstur: are you around? | 16:53 |
| *** mtanino has joined #openstack-manila | 16:53 | |
| vponomaryov | markstur: reaction on https://review.openstack.org/#/c/427811/ ? | 16:54 |
| vponomaryov | bswartz, gouthamr: what do we do with this ^ ? | 16:54 |
| vponomaryov | requested change by gouthamr is here -> https://review.openstack.org/#/c/428748/ | 16:55 |
| vponomaryov | it depends on above one | 16:55 |
| markstur | vponomaryov: I'm looking at it | 16:55 |
| tbarron | gouthamr: fix for https://review.openstack.org/#/c/428771/ looks good but shouldn't the commit message mention the cleanup failures and that they are causing sporadic tempest failures despite the test passing for the wrong reason? | 17:04 |
| tbarron | gouthamr: when I look a git log I like to be able to see what problems were fixed/when | 17:04 |
| *** chlong has quit IRC | 17:16 | |
| *** timcl has quit IRC | 17:19 | |
| *** xyang1 has quit IRC | 17:24 | |
| vponomaryov | bswartz: https://bugs.launchpad.net/manila/+bug/1661266 is not actually for RC1, we agreed to add "partial" fix,which is addon of DB model attr | 17:24 |
| openstack | Launchpad bug 1661266 in Manila "Share group types have no public extra specs" [Undecided,In progress] - Assigned to Valeriy Ponomaryov (vponomaryov) | 17:24 |
| vponomaryov | bswartz: all other stuff is expected to be backported later | 17:24 |
| vponomaryov | if added at all | 17:24 |
| *** timcl has joined #openstack-manila | 17:25 | |
| *** xyang1 has joined #openstack-manila | 17:25 | |
| openstackgerrit | Merged openstack/manila master: HNAS: Fix syntax to make shares read-only in snapshot create https://review.openstack.org/428267 | 17:27 |
| *** jcsp has quit IRC | 17:28 | |
| *** jcsp has joined #openstack-manila | 17:28 | |
| gouthamr | tbarron: yes.. sure thing. will do | 17:30 |
| tbarron | gouthamr: ty | 17:31 |
| openstackgerrit | Merged openstack/manila master: Fix HNAS driver inconsistent exceptions https://review.openstack.org/427660 | 17:32 |
| openstackgerrit | Goutham Pacha Ravi proposed openstack/manila master: Fix wrong access-rule negative test https://review.openstack.org/428771 | 17:32 |
| openstackgerrit | Yogesh proposed openstack/manila master: Reset the dummy driver method delay to zero https://review.openstack.org/426909 | 17:36 |
| openstackgerrit | Yogesh proposed openstack/manila master: Improve test coverage for share migration https://review.openstack.org/418559 | 17:37 |
| gouthamr | vponomaryov: why do you need to bump the API microversion for the key changes that you're not making right now? | 17:39 |
| gouthamr | vponomaryov: share groups only recently merged.. and the API is experimental... | 17:40 |
| gouthamr | vponomaryov: i'm okay with holding off till pike, but it makes any downstream testing harder | 17:40 |
| vponomaryov | gouthamr: we bump version for any API change | 17:40 |
| vponomaryov | gouthamr: no matter whether it is experimental or not | 17:41 |
| gouthamr | vponomaryov: hmmm, not so true. we've taken the liberty to avoid micro-version bumps in the past.. also, there's been only one (unrelated) API that was added between 2.31 and 2.32 | 17:42 |
| vponomaryov | gouthamr: http://docs.openstack.org/developer/manila/devref/api_microversion_dev.html#when-do-i-need-a-new-microversion | 17:42 |
| gouthamr | vponomaryov: my point is, it's not being consumed yet and we found a bug early.. maybe fix the bug without a microversion change.. because we don't want to advertise two different behaviors for a wrong | 17:43 |
| vponomaryov | gouthamr: sneaky proposal ) | 17:43 |
| gouthamr | vponomaryov: not at all.. experimental is sneaky.. but let's be consistent with how we treat these APIs | 17:43 |
| vponomaryov | gouthamr: anyway, we disable it for Ocata | 17:44 |
| bswartz | vponomaryov: the partial fix is for rc1 -- after that merges we can retarget bug out of ocata | 17:44 |
| gouthamr | vponomaryov: yeah, with an option of enabling.. | 17:45 |
| gouthamr | vponomaryov: i'm only asking for the key to be added so we can facilitate testing this in ocata.. | 17:45 |
| bswartz | vponomaryov: it has to be targeted to RC1 so we don't accidentally cut the milestone without that partial fix | 17:45 |
| vponomaryov | bswartz: ok | 17:45 |
| vponomaryov | gouthamr: if we do not add it to API, no need to change API tests | 17:46 |
| vponomaryov | gouthamr: Db attr is needed for some drivers that fail on snapshot deletion without it | 17:47 |
| vponomaryov | gouthamr: so, driver creates it for itself | 17:47 |
| vponomaryov | gouthamr: in scope of Ocata | 17:47 |
| gouthamr | vponomaryov: yes, i agree.. don't want to cause more work atm.. the bugfix you proposed solves the problem.. but as you know we do downstream testing with tempest too.. and we need the keys in the API to be sure.. again, not a biggie.. requesting you to consider it. | 17:48 |
| vponomaryov | gouthamr: I understand your point. Your testing will be "working existing tests on some backends" | 17:49 |
| vponomaryov | gouthamr: that were failing before | 17:49 |
| vponomaryov | gouthamr: like HNAS | 17:49 |
| vponomaryov | ganso: would be good to have testing results of https://review.openstack.org/#/c/427811/ with HNAS backend and share groups tests | 17:50 |
| gouthamr | vponomaryov: sure.. no regression while we are at this. | 17:51 |
| gouthamr | bbiab. will help review your patches when back | 17:51 |
| *** lpetrut has joined #openstack-manila | 17:54 | |
| *** mkoderer___ has quit IRC | 17:56 | |
| *** mkoderer__ has joined #openstack-manila | 17:56 | |
| *** andreaf has quit IRC | 17:56 | |
| *** smcginnis has quit IRC | 17:56 | |
| *** scottda has quit IRC | 17:56 | |
| *** smcginnis has joined #openstack-manila | 17:56 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Disable share groups APIs by default https://review.openstack.org/428840 | 17:57 |
| bswartz | thanks ganso I'm looking now | 17:58 |
| *** andreaf has joined #openstack-manila | 17:58 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Disable share groups APIs by default https://review.openstack.org/428840 | 17:59 |
| bswartz | tbarron: updating the tempest pin can happen any time, but more frequent is better than less frequent and not around release time is better than around release time | 17:59 |
| *** DuncanT has quit IRC | 17:59 | |
| bswartz | tbarron: given that we're getting close to a releast now, sooner is better than later | 18:00 |
| *** DuncanT has joined #openstack-manila | 18:00 | |
| * bswartz is pissed off at gerrit | 18:02 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Disable share groups APIs by default https://review.openstack.org/428840 | 18:02 |
| * bswartz throws a chair at openstackgerrit | 18:03 | |
| *** scottda has joined #openstack-manila | 18:04 | |
| vponomaryov | bswartz: what is wrong with gerrit? | 18:04 |
| vponomaryov | bswartz: it works just fine for me | 18:04 |
| bswartz | the web UI doesn't allow me to do online edit and read other's comments at the same time | 18:04 |
| bswartz | so I have to do old fashioned editing in my text editor | 18:05 |
| vponomaryov | bswartz: that is why it is useful to get used to use console editor always | 18:07 |
| bswartz | I use vi and pycharm most of the time | 18:07 |
| bswartz | for nitpicks I like to use gerrit web UI | 18:07 |
| tbarron | bswartz: I'll probably try two reviews then: one now at the latest commit, another after tempest team releases 15 | 18:11 |
| *** a-pugachev has quit IRC | 18:13 | |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: Pass access rules to driver on snapshot revert https://review.openstack.org/428398 | 18:14 |
| bswartz | tbarron: 2 updates? | 18:16 |
| bswartz | tbarron: that would be more than okay | 18:16 |
| bswartz | tbarron: I wouldn't mind bumping the commit every week or every day if it wasn't a drain on the team | 18:17 |
| tbarron | bswartz: ok, cool | 18:17 |
| bswartz | tbarron: the key is that we fix any issues that arise with each bump | 18:18 |
| bswartz | tbarron: the more often we do the bumps the less issues should need fixing | 18:18 |
| tbarron | bswartz: right | 18:18 |
| bswartz | why is devstack broken??? | 18:19 |
| bswartz | oh I think I know what's going on | 18:20 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Fix default approach for share group snapshot creation https://review.openstack.org/427811 | 18:20 |
| tbarron | bswartz: are you running out of memory? | 18:21 |
| vponomaryov | tbarron: look at gates )) | 18:22 |
| vponomaryov | tbarron: happy weekend! | 18:22 |
| vponomaryov | ^_^ | 18:22 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Add 'consistent_snapshot_support' attr to 'share_groups' DB model https://review.openstack.org/428748 | 18:23 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Disable share groups APIs by default https://review.openstack.org/428840 | 18:27 |
| vponomaryov | tbarron: hm, maybe false alert ) | 18:28 |
| ganso | bswartz: I am back now | 18:29 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix nonsense variable name https://review.openstack.org/428707 | 18:34 |
| ganso | bswartz: LGTM now, needs coverage improved only | 18:36 |
| bswartz | ganso: awesome | 18:37 |
| bswartz | tbarron: no it's my pip cache is hosed again | 18:38 |
| bswartz | tbarron: for some reason devpi just chokes sometimes | 18:38 |
| *** chlong has joined #openstack-manila | 18:38 | |
| bswartz | tbarron: and I don't find out about it until I attempt to restack | 18:39 |
| tbarron | bswartz: k, gotcha | 18:39 |
| ganso | vponomaryov: Valeriy, I replied to your comments in https://review.openstack.org/#/c/428526/ | 18:44 |
| bswartz | ganso: your recommended fix didn't turn out to work | 18:53 |
| bswartz | ganso: I finally was able to restack and run unit tests and they fail | 18:53 |
| bswartz | ganso: share_instance['id'] -- KeyError: 'id' | 18:53 |
| bswartz | ganso: I'm pretty sure we have to go back to the DB to get the instances which is why I'll have to move the access rule loading down into the child methods | 18:55 |
| ganso | bswartz: https://github.com/openstack/manila/blob/master/manila/db/sqlalchemy/models.py#L750 | 18:55 |
| ganso | bswartz: you'd need to debug to see what's wrong, the code that allows that ID to exist is here ^ | 18:56 |
| bswartz | xyang1 xyang_ vponomaryov cknight toabctl tbarron: still curious if you have any issues with: http://imgur.com/a/Thb5f | 18:56 |
| bswartz | ganso: does that thing have an "id" field? | 18:57 |
| xyang1 | bswartz: that's manila mascot? what is it | 18:58 |
| ganso | bswartz: it is supposed to https://github.com/openstack/manila/blob/master/manila/db/sqlalchemy/models.py#L361 | 18:58 |
| bswartz | xyang1: zorilla | 18:58 |
| ganso | xyang1: zorilla | 18:58 |
| openstackgerrit | Tom Barron proposed openstack/manila master: Update our tempest pin https://review.openstack.org/428230 | 18:58 |
| xyang1 | bswartz: not my favoriate. but if that's what we voted for, I am ok | 18:59 |
| bswartz | xyang1: it's not a skunk at least | 19:00 |
| bswartz | ganso: I think I see the problem | 19:00 |
| tbarron | seems innocuous enough | 19:00 |
| xyang1 | bswartz: but that is what wiki says:) | 19:00 |
| xyang1 | bswartz: African skunk | 19:00 |
| bswartz | the wiki says not a skunk | 19:00 |
| *** yumiriam has quit IRC | 19:00 | |
| bswartz | it may be called that but it's not related | 19:01 |
| xyang1 | bswartz: a type of weasel | 19:01 |
| tbarron | genotype/phenotype | 19:01 |
| xyang1 | bswartz: not a good choice for the year of rooster:) | 19:02 |
| tbarron | do zorilla's eat roosters? | 19:03 |
| ganso | bswartz: I sent the pic to my team, the whole team thought it was a skunk | 19:03 |
| tbarron | http://www.urbandictionary.com/define.php?term=zorilla | 19:03 |
| xyang1 | tbarron: weasel eats chicken | 19:03 |
| tbarron | xyang1: but roosters are tough chickens | 19:04 |
| xyang1 | tbarron: my friends sent me a drawing of a weasel saying happy new year to a group of frightened chicken | 19:04 |
| tbarron | xyang1: rofl | 19:04 |
| xyang1 | tbarron: it means "bad intentions" in chinese | 19:05 |
| tbarron | travelchinaguide.com says that "Rooster is almost the epitome of fidelity and punctuality" ? | 19:05 |
| tbarron | xyang1: I get the punctuality thing, but the first? | 19:05 |
| *** xyang_ has quit IRC | 19:06 | |
| tbarron | xyang1: weasels have a reputation world over for bad intentions I think | 19:06 |
| xyang1 | tbarron: so weasels are bad then | 19:06 |
| tbarron | but maybe a zorilla is special; none of us really know what it is I think. | 19:06 |
| gouthamr | skunk ^ | 19:06 |
| bswartz | weasels are adorable however | 19:06 |
| tbarron | That's an advantage: we can give it whatever reputation we want. | 19:06 |
| xyang1 | tbarron: wiki says it belongs to the family of weasels | 19:06 |
| tbarron | xyang1: don't judge everyone by their family :D | 19:07 |
| *** xyang_ has joined #openstack-manila | 19:07 | |
| bswartz | when you don't want the negative connotation of weasel you call them ferrets | 19:07 |
| xyang1 | tbarron: :) | 19:07 |
| gouthamr | tbarron: roosters are evil, and terrible | 19:07 |
| bswartz | or zorilla | 19:07 |
| tbarron | gouthamr: bad year to say that, 1.4B people will disagree with you | 19:08 |
| xyang1 | gouthamr: who says roosters are evil? this is the year of rooster. it has to be good | 19:08 |
| * tbarron can offer diplomatic advice all around to people who can't count well | 19:09 | |
| gouthamr | xyang1: well, i'd love 2022 :) | 19:09 |
| gouthamr | hahahaha tbarron | 19:09 |
| gouthamr | skunks though. fascinating animals.. | 19:10 |
| bswartz | ganso, gouthamr: https://github.com/openstack/manila/blob/master/manila/tests/fake_share.py#L154 | 19:11 |
| *** chlong has quit IRC | 19:11 | |
| bswartz | ganso: this seems to be a bug if the actual key for the share instance id is just "id" | 19:11 |
| gouthamr | bswartz: so... isn't snapshot_instance's instance's share_id the share instance ID? | 19:11 |
| bswartz | oh yes it has both | 19:12 |
| bswartz | so the share_id is the right key for the ID of the share, but there is no ID for the share instance in this fake | 19:12 |
| bswartz | so that's the bug | 19:12 |
| *** xyang_ has quit IRC | 19:13 | |
| ganso | bswartz: bug was only in test? | 19:13 |
| ganso | bswartz: *unit test | 19:13 |
| bswartz | ganso: IDK unit tests are all I've run so far | 19:13 |
| ganso | bswartz: I am waiting to see jenkins | 19:14 |
| ganso | bswartz: and my CI | 19:14 |
| *** crushil has quit IRC | 19:15 | |
| markstur | I thought the mascot thing was settled months ago when we lost our gorilla | 19:15 |
| ganso | bswartz: if I get home before I can test in my backend | 19:15 |
| bswartz | ganso new patchset coming | 19:15 |
| markstur | We could keep the logo and still call it a Manila Mink (maybe have coats made -- totally non-controversial) | 19:16 |
| markstur | besides -- everything eats chickens | 19:16 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: Pass access rules to driver on snapshot revert https://review.openstack.org/428398 | 19:17 |
| bswartz | ganso: we voluntarily dropped gorilla for a variety of reasons | 19:18 |
| bswartz | s/ganso/markstur/ | 19:18 |
| markstur | yeah I know, but still. Was an easy choice. If everyone would just stay off the internet. | 19:19 |
| *** eharney has quit IRC | 19:19 | |
| *** chlong has joined #openstack-manila | 19:21 | |
| *** chlong has quit IRC | 19:27 | |
| *** akapil has joined #openstack-manila | 19:38 | |
| *** akapil has quit IRC | 19:39 | |
| *** chlong has joined #openstack-manila | 19:40 | |
| *** xyang_ has joined #openstack-manila | 19:52 | |
| bswartz | ganso: this latest change seems to fail | 19:56 |
| *** xyang_ has quit IRC | 20:00 | |
| ganso | bswartz: looking | 20:03 |
| bswartz | ganso: I'm not sure why yet either | 20:06 |
| bswartz | but access rules are vanishing | 20:06 |
| *** dustins has quit IRC | 20:07 | |
| *** xyang_ has joined #openstack-manila | 20:08 | |
| bswartz | ganso: the problem I'm observing is that the snapshot LV isn't getting deleted as expected | 20:19 |
| bswartz | It could be a timing problem or some other error that's not appearing in the logs | 20:19 |
| bswartz | ganso: but for me I'm getting exceptions thrown at the step where it tries to recreate the snapshot LV | 20:19 |
| bswartz | it says it already exists | 20:20 |
| *** jprovazn has quit IRC | 20:21 | |
| *** xyang_ has quit IRC | 20:22 | |
| *** xyang_ has joined #openstack-manila | 20:26 | |
| *** draynium has quit IRC | 20:26 | |
| *** xyang_ has quit IRC | 20:27 | |
| *** xyang_ has joined #openstack-manila | 20:28 | |
| ganso | bswartz: you haven't changed anything related to the LVM driver in this latest patch | 20:36 |
| ganso | bswartz: didn't the previous one work? | 20:36 |
| ganso | bswartz: does it only happen when you have users mounted? I see PS #5 LVM passed in jenkins | 20:36 |
| openstackgerrit | Andreas Jaeger proposed openstack/manila master: Prepare for using standard python tests https://review.openstack.org/429032 | 20:40 |
| *** dustins has joined #openstack-manila | 20:44 | |
| *** xyang_ has quit IRC | 20:45 | |
| *** eharney has joined #openstack-manila | 20:51 | |
| ganso | gouthamr: if you could please review https://review.openstack.org/#/c/428526 when you have time | 20:52 |
| *** xyang_ has joined #openstack-manila | 20:54 | |
| *** a-pugachev has joined #openstack-manila | 20:56 | |
| *** xyang_ has quit IRC | 20:58 | |
| *** gouthamr has quit IRC | 20:59 | |
| *** xyang_ has joined #openstack-manila | 21:00 | |
| *** gouthamr has joined #openstack-manila | 21:05 | |
| *** catintheroof has quit IRC | 21:11 | |
| bswartz | ganso: really? | 21:17 |
| bswartz | ganso: I see unit tests failures on PS5, that's why I made PS6 | 21:18 |
| ganso | bswartz: I meant LVM job passed | 21:18 |
| ganso | bswartz: sorry | 21:18 |
| bswartz | ganso: yes it's able to pass the LVM functional test | 21:18 |
| bswartz | ganso: when I run my functional test over and over I begin to see failures over time | 21:19 |
| bswartz | ganso: it could be a timing problem | 21:19 |
| bswartz | ganso: or possibly corrupted state when I abort an operation in progress | 21:19 |
| ganso | bswartz: that is more likely | 21:20 |
| gouthamr | ganso: done. | 21:20 |
| ganso | gouthamr: thanks | 21:22 |
| gouthamr | ganso: did you see this: https://review.openstack.org/#/c/429032/1/tools/test-setup.sh | 21:22 |
| ganso | gouthamr: hadn't seen it! | 21:23 |
| ganso | gouthamr: looks handy! | 21:23 |
| gouthamr | ganso: yeah, couple of #TILs in there | 21:23 |
| bswartz | ganso: something somewhere isn't waiting long enough | 21:23 |
| bswartz | when I sleep for 20 seconds the tests passes every time | 21:23 |
| bswartz | when I sleep for 5 seconds it blows up pretty quick | 21:24 |
| ganso | bswartz: O_O | 21:24 |
| ganso | bswartz: shouldn't have to sleep | 21:24 |
| bswartz | ganso: I'm not sure if lvchange -an or lvchange -ay are REALLY done when those commands return | 21:25 |
| bswartz | and if they are not, we need a wait loop in the driver | 21:25 |
| *** crushil has joined #openstack-manila | 21:26 | |
| bswartz | ganso: are there any lockouts between access-allow and snapshot-create? | 21:27 |
| bswartz | it seems like the length of time I wait between access-allow and snapshot-create is the problem | 21:28 |
| *** crushil has quit IRC | 21:35 | |
| bswartz | gah! | 21:38 |
| bswartz | I have more than 1 problem in this dev env | 21:38 |
| bswartz | ganso anyways I'll keep working on this over the weekend | 21:38 |
| *** adrianofr_ has quit IRC | 21:39 | |
| *** chlong has quit IRC | 21:44 | |
| *** lpetrut has quit IRC | 21:47 | |
| *** xyang_ has quit IRC | 21:52 | |
| *** xyang_ has joined #openstack-manila | 21:56 | |
| *** gouthamr has quit IRC | 21:57 | |
| rhagarty | just grabbed latest and am seeing this error trying to create a share network from manila_ui - "got an unexpected keyword argument 'nova_net_id'". Any ideas? | 21:57 |
| dustins | rhagarty: When you attempted to create the share network, did you specify a nova network or a Neutron network? | 21:58 |
| *** kambiz has joined #openstack-manila | 21:59 | |
| dustins | I'm not a Horizon expert, but I'm willing to play one on TV :) | 21:59 |
| dustins | My first instinct is that a parameter is getting put in a place it's not supposed to be behind the scenes | 22:00 |
| rhagarty | dustins: not sure... I'm just using the default "private". I just realized I didn't update to the latest UI, maybe I need to update that first. | 22:01 |
| rhagarty | (I just update the service and client) | 22:01 |
| dustins | rhagarty: Yeah, give Horizon an update and see if that helps out | 22:01 |
| dustins | Which distribution are you using? | 22:01 |
| rhagarty | devstack - latest | 22:02 |
| dustins | Okay, awesome | 22:02 |
| dustins | IIRC, that private network that Devstack sets up is a Neutron network, so it's possible that the neutron network is being passed to the API as a nova network ID | 22:03 |
| dustins | It's been a little while since I've messed with Devstack | 22:03 |
| rhagarty | dustins: ok - let me update and check again... thanks | 22:04 |
| dustins | rhagarty: No worries! Good luck! | 22:04 |
| dustins | If it looks like something's up with it still, I'd open up a lauchpad bug so we can keep track of it | 22:04 |
| *** xyang_ has quit IRC | 22:05 | |
| dustins | I'm about to sign off soon for the weekend, but ping me in the beginning of next week if you have questions! | 22:05 |
| rhagarty | dustins: will do... thanks again | 22:06 |
| *** xyang_ has joined #openstack-manila | 22:06 | |
| *** Yogi1 has quit IRC | 22:11 | |
| *** xyang_ has quit IRC | 22:20 | |
| *** breitz has quit IRC | 22:26 | |
| *** breitz has joined #openstack-manila | 22:27 | |
| *** xyang_ has joined #openstack-manila | 22:29 | |
| *** xyang_ has quit IRC | 22:34 | |
| *** xyang_ has joined #openstack-manila | 22:35 | |
| *** dustins has quit IRC | 22:36 | |
| *** xyang_ has quit IRC | 22:39 | |
| openstackgerrit | Merged openstack/manila master: Fix wrong access-rule negative test https://review.openstack.org/428771 | 22:39 |
| *** xyang_ has joined #openstack-manila | 22:42 | |
| *** chlong has joined #openstack-manila | 22:42 | |
| *** xyang_ has quit IRC | 22:45 | |
| *** gouthamr has joined #openstack-manila | 22:45 | |
| openstackgerrit | Merged openstack/manila master: Fix migration of mountable snapshots https://review.openstack.org/428526 | 22:56 |
| *** xyang_ has joined #openstack-manila | 23:02 | |
| *** xyang_ has quit IRC | 23:20 | |
| *** xyang_ has joined #openstack-manila | 23:21 | |
| *** xyang_ has quit IRC | 23:23 | |
| openstackgerrit | Alex Schultz proposed openstack/puppet-manila master: Prepare ocata-rc1 release https://review.openstack.org/429089 | 23:23 |
| *** xyang1 has quit IRC | 23:31 | |
| bswartz | rhagarty: nova_net_id has been deprecated forever -- if the UI or client is still sending it, it's a bug | 23:32 |
| *** dsariel has quit IRC | 23:38 | |
| rhagarty | bswartz: I'm sure it's my issue... trying to rebase and test again. Thanks | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!