Thursday, 2020-03-19

« Wednesday, 2020-03-18 Index Friday, 2020-03-20 »
zaitcevJust use 800:46
*** ianychoi_ has joined #openstack-swift01:38
*** ianychoi has quit IRC01:40
*** renich has quit IRC02:00
*** psachin has joined #openstack-swift02:30
rledisezclayg: I did a small script to test compatibility between jerasure and isal. it works. did I miss something?02:34
rledisezhttp://paste.openstack.org/show/790862/02:34
rledisezinteresting thing is that the encoding does not produce the same fragments depending on the ec_type, but they can all decode each others02:35
rledisezcauchy also works if you remove the (6, 12) combination (does not cauchy support more parity frags than data frags?)02:36
rledisezIt also works if you drop a random frag, but for rs_vand only (rs_cauchy fails the assertion when ec_type is different)02:45
openstackgerritPete Zaitcev proposed openstack/swift master: Let developers/operators add watchers to object audit (simplified)  https://review.opendev.org/70665302:45
timburkerledisez, that's passing all the frags on every decode, isn't it? i'd try something more like kota_ did for https://bugs.launchpad.net/liberasurecode/+bug/1639691 and have another loop like `for check_frags in combinations(frags, k):`03:01
openstackLaunchpad bug 1639691 in PyECLib "EC: Swift can return corrupted Data and be able to go data lost at isa_l_rs_vand policy with >=5 parities" [Undecided,In progress]03:01
timburkegets expensive as hell, though, as the number of frags goes up. probably only want to look at the policies actually in-use!03:03
*** gyee has quit IRC04:08
*** baojg has quit IRC04:42
kota_rledisez: AFAIK, liberasurecode doesn't try to decode if all data fragments available in the given data set.05:08
kota_your script feeds all encoded fragments so that liberasurecode just concat all data fragments that should be same data payloads even the encode driver would be different.05:10
kota_i.e. if you feeds the subset of fragments list including parity frag, the decoding will fail as you expected.05:11
*** evrardjp has quit IRC05:36
*** evrardjp has joined #openstack-swift05:36
openstackgerritTim Burke proposed openstack/swift master: encryption: Expose decrypted metadata via CORS  https://review.opendev.org/71223706:01
*** tesseract has joined #openstack-swift07:51
*** threestrands has quit IRC08:01
*** pawan-gupta has quit IRC08:08
*** clayg has quit IRC08:08
*** pawan-gupta has joined #openstack-swift08:08
*** fyx has quit IRC08:09
*** clayg has joined #openstack-swift08:09
*** ChanServ sets mode: +v clayg08:09
*** fyx has joined #openstack-swift08:11
*** rpittau|afk is now known as rpittau08:11
*** tkajinam has quit IRC08:12
*** psachin has quit IRC09:01
*** psachin has joined #openstack-swift09:02
*** tesseract has quit IRC10:36
*** tesseract has joined #openstack-swift10:42
*** tesseract has quit IRC10:46
*** tesseract has joined #openstack-swift10:48
*** tesseract-RH has joined #openstack-swift11:00
*** tesseract-RH has left #openstack-swift11:02
*** tesseract has quit IRC11:03
*** rpittau is now known as rpittau|bbl11:30
*** rdejoux has quit IRC11:34
*** ccamacho has quit IRC11:38
*** ccamacho has joined #openstack-swift11:44
*** zaitcev has quit IRC11:55
*** rdejoux has joined #openstack-swift11:58
*** zaitcev has joined #openstack-swift12:29
*** ChanServ sets mode: +v zaitcev12:29
*** rpittau|bbl is now known as rpittau12:59
*** btorch has joined #openstack-swift13:16
*** psachin has quit IRC13:21
*** btorch has quit IRC13:46
*** btorch has joined #openstack-swift13:47
*** StevenK has quit IRC13:49
*** StevenK has joined #openstack-swift13:50
*** btorch has quit IRC14:19
*** btorch has joined #openstack-swift14:19
*** btorch has quit IRC15:02
*** btorch has joined #openstack-swift15:02
*** jvisser has quit IRC15:09
clayg@timburke https://usercontent.irccloud-cdn.com/file/rPEIJLBY/cors.png15:14
claygI think tr lightgreat is better than the nth child - but only marginally so 🤷‍♂️15:15
openstackgerritClay Gerrard proposed openstack/swift master: Move cors out of functests  https://review.opendev.org/71375415:18
*** btorch has quit IRC15:26
*** btorch has joined #openstack-swift15:26
*** jv has joined #openstack-swift15:28
zaitcevIs Vancouver PTG canceled?15:37
alecuyerzaitcev:  hello - I think it's supposed to be "virtual" now: http://lists.openstack.org/pipermail/foundation/2020-March/002854.html15:39
zaitcevalecuyer: thanks15:40
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: In _get_vfile(), handle the case when a referenced object has been punched  https://review.opendev.org/71390915:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Fix get_stats  https://review.opendev.org/71391015:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Raise an exception if get_vfile() fails, rather than ignoring the error  https://review.opendev.org/71391115:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Fix RPC error checking in losf volume check  https://review.opendev.org/71391215:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Make a new timeseries for LOSF object header mismatches  https://review.opendev.org/71391315:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Add a sleep() between index-server health checks  https://review.opendev.org/71391415:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Add missing option in for the volume check script  https://review.opendev.org/71391515:47
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Remove migrator.py  https://review.opendev.org/71391615:47
alecuyersorry about that15:48
rledisezjust upstream catching up with our internal branch ;)15:48
timburkealecuyer, nothing to apologize for :D16:40
timburkeclayg, fwiw, those 0 != 206 failures indicate that the CORS request was blocked, most likely because of a problem with origins16:43
*** btorch has quit IRC16:43
*** btorch has joined #openstack-swift16:43
*** gilles_biannic has joined #openstack-swift16:50
gilles_biannicHello. I see you're working on CORS subject :)16:52
gilles_biannicI have some clients complaining about cors troubles. However, each time I check by myself, it's working fine.16:55
gilles_biannicWhat they reported to me is :16:56
gilles_biannicThe chrome log contains "Access to image https://storage.bhs.cloud.ovh.net/....../image.png from origin https://website.com has been blocked by CORS policy: No ‘Access-Control-Allow-Origin’ header is present on the requested resource".16:57
gilles_biannicThe container has : "Meta Access-Control-Allow-Origin: *"16:57
gilles_biannicAnd they said that reloading the page made the issue disappear for them.16:58
gilles_biannicDoes that ring a bell to you ?17:01
*** btorch has quit IRC17:19
*** gyee has joined #openstack-swift17:19
*** btorch has joined #openstack-swift17:19
*** evrardjp has quit IRC17:36
*** evrardjp has joined #openstack-swift17:36
clayg"reloading the page made the issue disappear" 🤯17:40
claygsounds like a caching issue somewhere 😞17:40
claygit could be in the browser or a hop along the way (cdn?) - but it could also be in the cluster and memcache, @timburke is teaching me some of these values are set on the container17:41
timburkemakes me think of https://github.com/openstack/swift/commit/3e4607954671dee21107e6083f0695076987cc3e17:43
*** btorch has quit IRC17:43
*** btorch has joined #openstack-swift17:44
timburkebut with "Access-Control-Allow-Origin: *", it shouldn't really be an issue...17:45
timburkei agree, though -- sounds like a caching issue. gilles_biannic, do you know what kind of request they're making? is it a CORS "Simple Request", or does it require a preflight OPTIONS request?17:48
*** gilles_biannic has quit IRC17:55
timburkeclayg, not digging the alternating row colors? i'm fine with keeping it the old way -- i just thought you'd like the clear banding17:59
*** rpittau is now known as rpittau|afk18:06
*** btorch has quit IRC18:14
*** btorch has joined #openstack-swift18:15
*** btorch has quit IRC18:39
*** btorch has joined #openstack-swift18:40
*** gyee has quit IRC18:44
*** gyee has joined #openstack-swift18:51
*** btorch has quit IRC18:57
*** btorch has joined #openstack-swift18:58
*** btorch has quit IRC19:03
*** btorch has joined #openstack-swift19:03
*** btorch has quit IRC19:27
*** btorch has joined #openstack-swift19:27
*** gyee has quit IRC20:03
openstackgerritTim Burke proposed openstack/swift master: Add some functional CORS tests  https://review.opendev.org/53302820:56
openstackgerritTim Burke proposed openstack/swift master: cors: Add Content-Length to default-exposed headers  https://review.opendev.org/71223820:56
timburkeso i'm thinking i'd actually rather *not* have these all in a chain if i can help it -- i think there's value in the new gate job, but after that i ought to fan out and reduce the number of patches between where we are now and having CORS support for S321:10
openstackgerritMerged openstack/swift feature/losf: In _get_vfile(), handle the case when a referenced object has been punched  https://review.opendev.org/71390921:17
openstackgerritTim Burke proposed openstack/swift master: Add some functional CORS tests  https://review.opendev.org/53302821:38
openstackgerritTim Burke proposed openstack/swift master: cors: Add Content-Length to default-exposed headers  https://review.opendev.org/71223821:38
*** gyee has joined #openstack-swift22:00
mattoliveraumorning22:04
*** btorch has quit IRC22:13
*** btorch has joined #openstack-swift22:14
*** zigo has quit IRC22:22
*** zigo has joined #openstack-swift22:32
*** btorch has quit IRC22:37
*** btorch has joined #openstack-swift22:38
*** btorch has quit IRC23:07
*** btorch has joined #openstack-swift23:08
« Wednesday, 2020-03-18 Index Friday, 2020-03-20 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!