Monday, 2020-10-12

*** zzzeek has quit IRC00:20
*** zzzeek has joined #openstack-cinder00:23
*** rcernin has quit IRC00:26
*** zenkuro has quit IRC00:50
*** rcernin has joined #openstack-cinder00:54
*** n-saito has quit IRC00:54
*** Liang__ has joined #openstack-cinder01:21
*** mpsairam has quit IRC01:27
*** pmannidi has joined #openstack-cinder01:27
*** ociuhandu has joined #openstack-cinder02:30
*** ociuhandu has quit IRC02:34
*** pmannidi has quit IRC03:03
*** udesale has joined #openstack-cinder03:05
*** Liang__ has quit IRC03:27
*** Liang__ has joined #openstack-cinder03:27
*** ircuser-1 has quit IRC03:49
*** psachin has joined #openstack-cinder03:51
*** pmannidi has joined #openstack-cinder04:19
*** manoj_kumar_kata has joined #openstack-cinder04:21
*** vishalmanchanda has joined #openstack-cinder04:39
*** psachin has quit IRC04:39
*** rcernin has quit IRC04:47
*** m75abrams has joined #openstack-cinder05:01
*** dsariel has joined #openstack-cinder05:02
*** abdysn has joined #openstack-cinder05:11
*** rcernin has joined #openstack-cinder05:18
*** Abdallahyas has joined #openstack-cinder05:32
*** abdysn has quit IRC05:36
*** Abdallahyas has quit IRC05:42
*** abdysn has joined #openstack-cinder05:42
*** Liang__ has quit IRC06:02
*** Liang__ has joined #openstack-cinder06:03
*** e0ne has joined #openstack-cinder07:01
*** tosky has joined #openstack-cinder07:03
*** zenkuro has joined #openstack-cinder07:09
*** zenkuro has quit IRC07:15
*** zenkuro has joined #openstack-cinder07:15
*** JaySee has quit IRC07:15
*** rcernin has quit IRC07:26
*** zenkuro has quit IRC07:30
*** zenkuro has joined #openstack-cinder07:30
*** sharathkacham has joined #openstack-cinder07:31
*** brinzhang has joined #openstack-cinder07:32
*** rpittau|afk is now known as rpittau07:35
*** benfelin has joined #openstack-cinder07:43
*** sharathkacham has quit IRC07:46
*** priteau has joined #openstack-cinder07:51
*** sharathkacham has joined #openstack-cinder07:53
*** ociuhandu has joined #openstack-cinder08:07
*** rcernin has joined #openstack-cinder08:10
*** e0ne has quit IRC08:22
*** e0ne_ has joined #openstack-cinder08:22
*** e0ne has joined #openstack-cinder08:23
*** e0ne_ has quit IRC08:23
*** rcernin has quit IRC08:33
*** brinzhang_ has joined #openstack-cinder08:33
*** brinzhang has quit IRC08:35
*** elod is now known as elod_pto08:41
zenkurohello guys, can you please review https://review.opendev.org/#/c/726906/, it has passed many code review and been approved by smcginnis and lseki. With last fix I removed redundant code and improved CI stability08:43
*** manoj_kumar_kata has quit IRC08:51
*** manoj_kumar_kata has joined #openstack-cinder08:51
*** dsariel has quit IRC09:09
*** zenkuro has quit IRC09:17
*** zenkuro has joined #openstack-cinder09:17
*** manoj_kumar_kata has quit IRC09:40
*** manoj_kumar_kata has joined #openstack-cinder09:40
*** dsariel has joined #openstack-cinder09:41
*** Abdallahyas has joined #openstack-cinder09:46
*** abdysn has quit IRC09:49
*** openstackgerrit has joined #openstack-cinder09:50
openstackgerritWenping Song proposed openstack/cinder master: Use TOX_CONSTRAINTS_FILE instead of obsolete UPPER_CONSTRAINTS_FILE  https://review.opendev.org/75752209:50
*** Liang__ has quit IRC10:09
openstackgerritfengjiankui proposed openstack/cinder master: create encryption volume with existing secret  https://review.opendev.org/75752810:21
openstackgerritfengjiankui proposed openstack/cinder master: create encryption volume with existing secret  https://review.opendev.org/75752810:25
*** pcaruana has quit IRC10:31
*** lpetrut has joined #openstack-cinder10:36
*** Luzi has joined #openstack-cinder10:52
*** pcaruana has joined #openstack-cinder11:03
*** udesale_ has joined #openstack-cinder11:04
*** udesale has quit IRC11:07
*** brinzhang_ has quit IRC11:37
*** brinzhang_ has joined #openstack-cinder11:37
zenkurolseki: hi, can you please give a blessing to a patch-set  https://review.opendev.org/#/c/726906/ once more?11:58
*** priteau has quit IRC11:59
zenkurosmcginnis: thanks a lot!12:08
smcginniszenkuro: No problem, thanks for getting all of the comments addressed. We should be in great shape now that wallaby development is open.12:10
zenkurosmcginnis: now when I have +2 there should be one more person to give +2? or something else should happen?(checking if I missing something)12:35
openstackgerritAjitha Robert proposed openstack/cinder master: Add support for reverting to snapshot in the Nimble Storage iSCSI and FC drivers  https://review.opendev.org/75755412:38
kmarceharney: I'm happy to see that cinder is going to be type checked now! :-) kudos!12:50
eharneykmarc: thanks, i'm looking forward to it :)12:51
*** GirishChilukuri has joined #openstack-cinder12:53
kmarcone tiny question, why the `import typing as ty` and ten use the "built-in" types as ty.Dict etc? I understand that import * is not favored and also that `import Dict,Whatever from typing` is a bit cumbersome, but I always do it like that; def fun() -> ty.Dict[...] looks a bit messy to me12:54
kmarcalthough, super polite! (ty == thank you)12:54
kmarc:-)12:54
eharneykmarc: i'm still a bit undecided on what style would be best, i've considered going that route too12:57
eharneyso, no strong reason really12:58
*** rosmaita has joined #openstack-cinder12:58
kmarcI'd guess IDEs like pycharm can do the auto-import, but TBH even using plain vim (without any fancy plugins to do this kinda stuff) was sufficient for me, although small project13:00
eharneynot sure, i stil mostly live in the vim world13:02
eharneyi'll try posting a patch with that style too for review and see what folks think13:04
openstackgerritEric Harney proposed openstack/cinder master: mypy: annotate volume manager (alternate style)  https://review.opendev.org/75756013:22
hemnaman that syntax is odd/ugly13:26
hemnare: type checking13:26
*** Luzi has quit IRC13:26
openstackgerritHelen Walsh proposed openstack/cinder stable/rocky: PowerMax Driver - Legacy volumes fail to live migrate  https://review.opendev.org/75756313:26
eharneyhemna: well, this patch may make it nicer ^13:26
eharneyhemna: but, any specific issues with it?13:26
hemnajust haven't seen it before13:28
hemnakinda clutters things up13:28
hemnathis might make backporting bug fixes to py2 code messy13:30
hemnawhy is there # type: ignore in some placess for self, and not others?13:32
hemnaseems inconsistent13:32
eharneyargs aren't currently matched between the Manager and VolumeManager classes, i didn't want to get into fixing all that right out of the gate13:33
eharneyit's as inconsistent as our code is currently :)13:33
*** pcaruana has quit IRC13:35
zenkurorosmaita: hi, if you have some free time can you please take a look at https://review.opendev.org/#/c/726906/, Ive addressed most of an issues and improved CI stability13:36
rosmaitazenkuro: i was just looking, but would rather wait for lseki to make sure all his concerns were addressed13:37
rosmaitabut if lseki tells me he is too busy today, i will review13:38
rosmaitazenkuro: you are talking about the new jovian dss driver, right? or am i confusing you with someone else13:38
zenkurorosmaita: yes, JovianDriver13:39
rosmaitaok13:39
zenkurorosmaita: thanks for your attention, I will wait for lseki then13:39
rosmaitazenkuro: feel free to ping me later if there is no action ... now that i think about it, it may be a holiday where lseki is today13:40
zenkurorosmaita: ok, will ping you later if nothing changes13:45
*** Abdallahyas has quit IRC14:05
*** TusharTgite has joined #openstack-cinder14:14
hemnaI would think type checking would be more useful between the volume manager and the drivers14:14
hemnathe volume manager hasn't changed much in a long time14:14
hemnawhat are we trying to gain by adding type checking now in the life of cinder?14:15
eharneyyes, it will be handy there for sure14:15
TusharTgitehi team14:16
hemnathe interface between the volume manager and the driver is where I think it might be useful14:16
TusharTgitePlease review the code changes https://review.opendev.org/#/c/749658/  need +2 workflow approvals to merge code changes.14:16
*** m75abrams has quit IRC14:20
noonedeadpunkhey everyone! I'm trying to understand why I have volume creation error when barbican is used14:20
noonedeadpunkhttp://paste.openstack.org/show/798953/14:20
noonedeadpunklog from barbican http://paste.openstack.org/show/798944/14:20
noonedeadpunkI'd say that it's barbican issue, but have one nuance...14:21
noonedeadpunkI've tried to raise here to see if store is even called here https://opendev.org/openstack/barbican/src/branch/master/barbican/plugin/resources.py#L71 but it's not14:21
redrobotnoonedeadpunk, looks like barbican is trying to decode some data that doesn't need decoding.14:22
noonedeadpunkso I'm under impression that we ask for a secret without trying to store it?14:22
noonedeadpunkredrobot: so, um, `b'\xe2\xca7\xc2i$\x99\x8a\xb5\xf6r\x9d&@\xe6\xf1\xfbW\xfeL\x1d\xae\xa1\x8e\x1b\xe6\xea\xaa\xc0\xf5\xd3\xa1'` doesn't need decoding?14:25
redrobotright, that is likely the actual secret14:25
redrobotnoonedeadpunk, len(b'\xe2\xca7\xc2i$\x99\x8a\xb5\xf6r\x9d&@\xe6\xf1\xfbW\xfeL\x1d\xae\xa1\x8e\x1b\xe6\xea\xaa\xc0\xf5\xd3\xa1') == 32, so that'14:26
redrobotthat's a 256 byte key14:26
noonedeadpunkhm, let me check that14:26
noonedeadpunkredrobot: you're right....14:30
noonedeadpunkadding here https://github.com/openstack/oslo.serialization/blob/master/oslo_serialization/base64.py#L68 try/except pass and volume is created14:31
redrobotnoonedeadpunk, base64 strings alwas end in "==", so it's easy to tell when the data is/isn't b64-encoded.14:34
redrobotHmm... it's strange to try to force ascii right there14:35
redrobotnoonedeadpunk, base64.b64decode() expects bytes and returns bytes, so I'm not sure I understand why Oslo is trying to force ascii encoding there14:41
redrobotgiven that ascii has very particular rules about what bytes are allowed.14:42
noonedeadpunkredrobot: maybe kind of left from py2?14:45
*** TusharTgite has quit IRC14:47
noonedeadpunkredrobot: are you going to patch it?14:48
lsekihi zenkuro, I'll take a look on your patch right now14:51
zenkurolseki: cool! thanks a lot!14:53
redrobotnoonedeadpunk, yeah, I can make a patch to oslo to remove that14:59
noonedeadpunkredrobot: well, don't really hurry. as once I dropped it another thing raised ....15:00
noonedeadpunkhttp://paste.openstack.org/show/798959/15:01
redrobotnoonedeadpunk, yeah, so that's gonna happen when you have something that is not encoded and you try to decode it15:01
noonedeadpunkbut um, `encoded = encoded.decode('ascii')` makes things work in such case15:02
redrobotSo, key == base64.decode(base64.encode(key)), but it seems that workflow is trying to do key == base64.decode(key) which is15:02
redrobotobviously not right15:03
noonedeadpunkit's totally not...15:06
redrobotthe decode('ascii') function takes a string that is an ascii string and turns it into the bytes15:07
redrobotbut the input HAS to be an ascii string15:07
redrobotbut in that workflow the input is just random bytes (ie, a key)15:07
redrobotso there's no way that .decode('ascii') is going to work 100% of the time15:07
GirishChilukurismcginnis lseki I have addressed review comments for code changes https://review.opendev.org/#/c/751994/, Please review the code changes for +2 approvals.15:07
redrobotit _might_ work some times, when every byte in key is in the ascii space15:07
redrobotbut as soo as the key as a non-ascii byte anywhere in the byte string you'll get that encoding error15:08
redrobot*as soon as15:08
noonedeadpunkredrobot: well yes, I titally agree with that, but don't see plain fix... as seems that just removal of decode to ascii is not really a solution. and adding except: pass sounds weird a bit as well:)15:09
noonedeadpunkand I don't see issue in barbican logic itself15:09
redrobotnoonedeadpunk, I'd have to look into the code doing the initial storage of the secret to be sure, but my guess is that:15:10
redrobotWhen doing the secret store, the code is sending payload=base64(key) and payload_content_type=base64.15:10
redrobotthis tells Barbican API that the thing you want to store is "key", but that it has been base64 encoded to be put into a json string.  (JSON does not support literals with arbitrary bytes, so base64 is necessary)15:11
redrobotso Barbican API does secret_payload=base64.decode(incoming_payload)15:12
redrobotand then later when you retrieve it from barbican the code is expecting base64(key) but barbican is only sending back key15:12
redrobotthe whole ascii thing in oslo seems bad to me.  especially because every string in Python 3 is UTF-8, not ascii.15:13
redrobot*payload_content_encoding=base6415:14
redrobotsorry, I typed the wrong param first15:14
*** manoj_kumar_kata has quit IRC15:15
noonedeadpunkwell yes, makes sense then15:16
noonedeadpunkbut yeah, ascii thing left from py2 I'm pretty sure about that... it's smth that sould be removed/replaced for sure15:16
rosmaitaredrobot: noonedeadpunk: i suggest writing some tests to prove this is an error before making any changes to oslo15:16
lsekiGirishChilukuri: sure, I' take a look15:16
rosmaitaredrobot: noonedeadpunk: we're dealing with a byte array in py3, so it has to be decoded to something before it can be interpreted as characters15:17
zenkurolseki: thanks for a feedback. I will start working on recommendations and improvements right after this patch pass15:19
redrobotrosmaita, I don't think that's accurate.  In the base64 module for the stdlib both encode and decode require byte strings as input and both output bytes as well.15:19
redrobotrosmaita, IIUC, the only time you need to decode something to pass it as a param to base64 is when you try to pass a string, in that case it has to be decoded from UTF-8 to get a byte array15:20
rosmaitaredrobot: i don't claim to be an expert on this, but since you are using the cinder channel, i figure i can butt in and request that if you patch this, you include a bunch of tests15:22
lsekizenkuro: awesome, thanks!15:22
redrobotrosmaita, understood, will definitely add tests around that15:22
rosmaita:D15:23
noonedeadpunkI guess it's my fault about helding dicsussion here:)15:23
redrobot:D15:23
rosmaitathis character encoding stuff is really tricky15:23
redrobotHeh, yes... we had to spend a lot of time looking into it to get the 2/3 compat stuff just right in barbican15:25
*** GirishChilukuri has quit IRC15:37
rosmaitasmcginnis: i am sorry i missed this earlier, but should we revert https://review.opendev.org/#/c/756456/ to remove the change to the sqlalchemy migration file?  i don't know that we want those different across releases unless absolutely necessary15:48
*** GirishChilukuri has joined #openstack-cinder15:49
*** rpittau is now known as rpittau|afk16:01
*** udesale_ has quit IRC16:16
*** _pewp_ has quit IRC16:16
*** _pewp_ has joined #openstack-cinder16:19
*** priteau has joined #openstack-cinder16:20
*** zenkuro has quit IRC16:20
*** tosky has quit IRC16:21
*** zenkuro has joined #openstack-cinder16:22
*** ociuhandu_ has joined #openstack-cinder16:35
*** ociuhandu has quit IRC16:39
*** ociuhandu_ has quit IRC16:40
*** zenkuro has quit IRC16:44
*** zenkuro has joined #openstack-cinder16:45
*** zenkuro has quit IRC16:46
*** zenkuro has joined #openstack-cinder16:46
*** zenkuro has quit IRC16:48
*** hamalq has joined #openstack-cinder16:49
*** sharathkacham has quit IRC16:55
openstackgerritAjitha Robert proposed openstack/cinder master: Nimble: Add support for revert to snapshot  https://review.opendev.org/75755417:05
*** pcaruana has joined #openstack-cinder17:13
*** zenkuro has joined #openstack-cinder17:19
zenkurohi, if I have a GATE failure which is not related to my code, what should I do https://zuul.opendev.org/t/openstack/build/c1f526b8cc924ca2af1990fd361e15d7 ?17:20
zenkuroin integrated storage17:21
*** zenkuro has quit IRC17:25
*** zenkuro has joined #openstack-cinder17:25
rosmaitazenkuro: recheck17:32
rosmaitawhich i see you already did17:32
zenkurorosmaita: yep, preventive recheck, Im asking since test will finish in 1.1 hour and you guys will go afk17:34
zenkurorosmaita: and Im not sure if recheck will help with gate job as well17:35
rosmaitazenkuro: yes, it should go to gate after check has passed17:36
rosmaitai will be around and will keep an eye on things17:36
zenkurorosmaita: \>_</17:36
*** ajitha has joined #openstack-cinder17:36
smcginnisajitha: No, release notes do not need to have a bug associated (better to talk in channel).17:50
ajithaok17:51
ajithai thought this would be the syntax,  tox -e venv -- reno new <bug-id>17:54
*** lpetrut has quit IRC17:59
GirishChilukuri@core reviewers Please review this code change https://review.opendev.org/#/c/744664/ for +2 approvals.18:10
*** GirishChilukuri has quit IRC18:15
*** zenkuro has quit IRC18:24
*** zenkuro has joined #openstack-cinder18:24
e0nehi. can anybody please help me with tools/config/generate_cinder_opts.py? it generates incorrect code for my change18:30
e0nehere is it's output: https://zuul.opendev.org/t/openstack/build/fb331539855048b8a48cb8e3c13bff8418:31
rosmaitae0ne: i am looking, but no progress to report yet18:59
e0nerosmaita: thanks18:59
ajithaHi all, i have submitted the revert to snapshot patch for review. in our CI we are still using devstack-gate, and included DEVSTACK_GATE_TEMPEST_REGEX parameter to include tempest tests.. how to include cinder_tempest_plugin tests with this parameter??19:01
smcginnise0ne: Did you run "tox -e genopts" for your patch?19:01
smcginnisHmm, does look messed up though.19:02
e0nesmcginnis: I didn't, because it generates invalid code:(19:02
*** priteau has quit IRC19:04
e0nesmcginnis: I don't understand what is going wrong with my patch19:04
lsekiyou may add an OR operator to the regex19:09
lsekisomething like `export DEVSTACK_GATE_TEMPEST_REGEX='tempest.api.volume|cinder_tempest_plugin'`19:10
lsekiajitha: ^^19:11
* smcginnis wants to rewrite generate_cinder_opts.py19:12
rosmaitayeah, a bit too much manual string matching19:13
smcginnisMy guess is that it doesn't anticipate the , at the end of line 102 here: https://review.opendev.org/#/c/722655/11/cinder/backup/drivers/ceph.py19:13
rosmaitayes, if you just have CONF.register_opts(CEPH_BACKUP_OPTS)  that first error goes away19:14
rosmaitathe [, problem still happens, though19:14
*** masterpe has quit IRC19:16
e0nerosmaita: thanks, -1 error is a good start :)19:19
smcginnisI think ideally this should just be redone to use something like I did with the sphinx extension: https://opendev.org/openstack/cinder/src/branch/master/doc/ext/driver_opts.py19:20
smcginnisAnd get rid of all this weird string matching and parsing and formatting.19:20
*** masterpe has joined #openstack-cinder19:23
rosmaitasmcginnis: e0ne: take a look at this: https://review.opendev.org/#/c/722655/11/cinder/volume/configuration.py@13719:24
e0nerosmaita: looking on it...19:26
rosmaitafeel free to tell me it's a stupid suggestion, on closer look i'm not so sure19:26
ajitha@lseki19:30
ajithathank you19:31
e0nerosmaita, smcginnis:  https://review.opendev.org/#/c/722655/11/cinder/volume/configuration.py@137 isn't related to genopts issue, but now I know where it is. thanks, for the help!19:34
e0neI'll try to fix it tomorrow19:35
rosmaita\o/19:35
e0nehere is it: https://review.opendev.org/#/c/722655/11/cinder/backup/manager.py@10119:36
*** benfelin has quit IRC19:49
*** e0ne has quit IRC19:50
*** e0ne has joined #openstack-cinder19:54
*** zenkuro has quit IRC19:56
*** zenkuro has joined #openstack-cinder19:57
*** e0ne has quit IRC19:59
*** zenkuro has quit IRC20:03
*** zenkuro has joined #openstack-cinder20:03
*** tosky has joined #openstack-cinder20:26
*** whoami-rajat__ has joined #openstack-cinder20:36
whoami-rajat__lseki: hi, around?20:36
lsekiwhoami-rajat__: o/20:38
whoami-rajat__lseki: does netapp nfs support qcow2 format volumes as well or only raw ?20:43
lsekiI think it supports them both, but let me check20:50
whoami-rajat__sure, if it supports qcow2 we might have a problem ...20:51
lsekiwhoami-rajat__: hmm qcow2 is converted to raw when I create a volume from a glance image, I need to try other ways to test it21:04
lsekiI'll try it tomorrow and let you know21:04
lsekiwhat's the problem?21:05
*** e0ne has joined #openstack-cinder21:09
whoami-rajat__lseki: the problem is related to glance cinder store, it's a long issue, i will detail you out tomorrow21:15
whoami-rajat__let me know if it supports qcow2 at any place21:16
*** vishalmanchanda has quit IRC21:27
*** e0ne has quit IRC21:27
*** rcernin has joined #openstack-cinder21:47
*** rcernin has quit IRC21:47
*** rcernin has joined #openstack-cinder21:47
openstackgerritBrian Rosmaita proposed openstack/cinder master: Preparing for removal of six.reraise()  https://review.opendev.org/75765521:54
*** rcernin has quit IRC22:05
*** ajitha has quit IRC22:14
*** rcernin has joined #openstack-cinder22:28
*** rcernin has quit IRC22:29
*** rcernin has joined #openstack-cinder22:30
*** watersj has quit IRC23:01
*** CeeMac has quit IRC23:01
*** tosky has quit IRC23:15
*** zenkuro has quit IRC23:37
*** hamalq has quit IRC23:52

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!