Monday, 2015-06-01

*** charlesw has joined #openstack-swift00:04
*** ho has joined #openstack-swift00:16
*** remix_tj has quit IRC00:17
*** remix_tj has joined #openstack-swift00:19
*** bi_fa_fu_ has joined #openstack-swift00:26
*** bi_fa_fu has quit IRC00:26
*** zynisch_o7 has joined #openstack-swift00:52
*** bi_fa_fu_ has quit IRC00:53
*** bi_fa_fu has joined #openstack-swift00:53
*** zynisch_o7 has quit IRC00:57
*** bi_fa_fu has quit IRC01:19
*** evanjfraser has joined #openstack-swift01:30
*** jrichli has joined #openstack-swift01:53
*** blmartin has joined #openstack-swift02:10
*** asettle is now known as asettle-afk02:21
*** zynisch_o7 has joined #openstack-swift02:41
*** zynisch_o7 has quit IRC02:46
*** blmartin has quit IRC02:51
openstackgerritHisashi Osanai proposed openstack/swift: Enable Object Replicator's failure count in recon  https://review.openstack.org/13834202:55
*** _bluev has quit IRC03:30
*** ppai has joined #openstack-swift03:43
*** haomaiw__ has joined #openstack-swift03:47
*** haomaiwang has quit IRC03:49
*** haomaiw__ has quit IRC03:59
*** haomaiwang has joined #openstack-swift04:02
*** charlesw has quit IRC04:13
*** jrichli has quit IRC04:18
*** asettle-afk is now known as asettle04:18
*** zynisch_o7 has joined #openstack-swift04:27
*** zynisch_o7 has quit IRC04:32
*** zhill_ has joined #openstack-swift04:46
*** zhill_ has quit IRC04:52
*** dimasot has joined #openstack-swift05:27
*** Trixboxer has joined #openstack-swift05:28
*** SkyRocknRoll has joined #openstack-swift05:47
*** SkyRocknRoll has joined #openstack-swift05:47
*** _bluev has joined #openstack-swift05:48
*** Triveni has joined #openstack-swift05:51
*** ppai has quit IRC06:01
*** zynisch_o7 has joined #openstack-swift06:16
*** zynisch_o7 has quit IRC06:21
*** _bluev has quit IRC06:21
cschwedeGood Morning!06:27
mattoliveraucschwede: morning!06:36
cschwedeHello Matthew!06:36
hocschwede: Morning!06:37
cschwedeHello Hisashi!06:38
*** SkyRocknRoll has quit IRC06:41
openstackgerritMichael Barton proposed openstack/swift: Don't quarantine on read_metadata ENOENT  https://review.openstack.org/18711606:46
openstackgerritChristian Schwede proposed openstack/swift: Make swift-recon compatible for servers without storage policies  https://review.openstack.org/18711907:03
*** ppai has joined #openstack-swift07:04
*** SkyRocknRoll has joined #openstack-swift07:09
*** [1]dimasot has joined #openstack-swift07:10
*** dimasot has quit IRC07:12
*** [1]dimasot is now known as dimasot07:12
*** ppai has quit IRC07:16
*** ppai has joined #openstack-swift07:23
*** haomaiw__ has joined #openstack-swift07:24
*** haomaiwang has quit IRC07:27
*** leopoldj has joined #openstack-swift07:32
*** geaaru has joined #openstack-swift07:33
openstackgerritMerged openstack/swift: Replaced setting run_pause with standard interval  https://review.openstack.org/17117007:37
*** jistr has joined #openstack-swift07:39
*** chlong has quit IRC07:45
mattoliverauI'm calling it a night, night all.07:46
cschwedemattoliverau: Have a nice evening! See ya07:49
*** acoles_away is now known as acoles07:55
acolesmattoliverau: g'night07:56
acolescschwede: good morning!07:56
cschwedeacoles: Good Morning Alistair!07:57
*** jordanP has joined #openstack-swift08:01
*** silor has joined #openstack-swift08:03
*** zynisch_o7 has joined #openstack-swift08:04
*** zynisch_o7 has quit IRC08:09
*** xnox has quit IRC08:10
*** xnox has joined #openstack-swift08:12
*** [1]dimasot has joined #openstack-swift08:14
*** dimasot has quit IRC08:17
*** [1]dimasot is now known as dimasot08:17
*** marzif_ has joined #openstack-swift08:21
*** krykowski has joined #openstack-swift08:23
*** admin6 has joined #openstack-swift08:30
*** hseipp has joined #openstack-swift08:30
*** hseipp has quit IRC08:31
*** hseipp has joined #openstack-swift08:31
*** yuan has quit IRC09:25
*** yuan has joined #openstack-swift09:25
*** aix has joined #openstack-swift09:43
*** zynisch_o7 has joined #openstack-swift09:53
*** zynisch_o7 has quit IRC09:58
*** openstackstatus has quit IRC10:20
*** openstack has quit IRC10:20
*** openstack has joined #openstack-swift10:27
*** foexle has joined #openstack-swift10:30
*** mmcardle has joined #openstack-swift10:38
*** [1]dimasot has joined #openstack-swift10:39
*** distroguy has joined #openstack-swift10:42
*** km__ has joined #openstack-swift10:42
*** Triveni has quit IRC10:43
*** treyd has quit IRC10:43
*** dimasot has quit IRC10:43
*** mitz has quit IRC10:43
*** km has quit IRC10:43
*** csmart has quit IRC10:43
*** mordred has quit IRC10:43
*** SkyRocknRoll has quit IRC10:43
*** mitz has joined #openstack-swift10:43
*** [1]dimasot is now known as dimasot10:43
*** mordred has joined #openstack-swift10:44
*** SkyRocknRoll has joined #openstack-swift10:45
*** treyd_ has joined #openstack-swift10:46
openstackgerritMark Seger proposed openstack/swift: Catch file open errors before calling ConfigParser  https://review.openstack.org/18677911:25
*** aix has quit IRC11:34
*** ppai has quit IRC11:36
*** zynisch_o7 has joined #openstack-swift11:42
*** zynisch_o7 has quit IRC11:47
*** mmcardle has quit IRC11:47
*** aix has joined #openstack-swift11:48
*** ppai has joined #openstack-swift11:48
*** admin6_ has joined #openstack-swift11:51
*** admin6 has quit IRC11:52
*** admin6_ is now known as admin611:52
*** ppai has quit IRC11:54
*** ppai has joined #openstack-swift12:02
*** SkyRocknRoll has quit IRC12:08
eikkenotmyname: thanks!12:09
*** HiramAbif has quit IRC12:11
*** kei_yama has quit IRC12:26
*** ho has quit IRC12:26
*** km__ has quit IRC12:27
*** SkyRocknRoll has joined #openstack-swift12:27
openstackgerritMark Seger proposed openstack/swift: Catch file open errors before calling ConfigParser  https://review.openstack.org/18677912:43
*** dimasot has quit IRC12:48
*** chlong has joined #openstack-swift12:49
*** mwheckmann has joined #openstack-swift12:49
*** bi_fa_fu has joined #openstack-swift12:54
*** ppai has quit IRC12:55
*** janonymous_ has joined #openstack-swift13:06
openstackgerritMerged openstack/swift: Make swift-recon compatible for servers without storage policies  https://review.openstack.org/18711913:10
*** SkyRocknRoll has quit IRC13:22
*** mmcardle has joined #openstack-swift13:22
*** annegentle has joined #openstack-swift13:23
*** acampbel11 has joined #openstack-swift13:32
*** acampbell has joined #openstack-swift13:32
*** B4rker has joined #openstack-swift13:34
*** charlesw has joined #openstack-swift13:58
*** blmartin has joined #openstack-swift13:59
*** annegentle has quit IRC14:02
*** annegentle has joined #openstack-swift14:03
*** janonymous_ has quit IRC14:03
*** jrichli has joined #openstack-swift14:04
*** petertr7_away is now known as petertr714:10
*** wbhuber has joined #openstack-swift14:27
openstackgerritPete Zaitcev proposed openstack/swift: Catch file open errors before calling ConfigParser  https://review.openstack.org/18677914:28
*** krykowski has quit IRC14:28
portantetorgomatic: regarding readv2/writev2, looks like patch is stalled, a few detractors upstream for it, need a solid use case for it, sounds like to me Swift would be a good one14:29
*** ptb has joined #openstack-swift14:32
*** annegentle has quit IRC14:32
*** breitz has joined #openstack-swift14:32
*** minwoob has joined #openstack-swift14:38
portantetorgomatic: based on my read of the patch set here: http://thread.gmane.org/gmane.linux.kernel.aio.general/4227/focus=430614:53
*** acampbell has quit IRC14:54
*** acampbel11 has quit IRC14:54
portantethe new readv2/writev2 system calls would require issuing a read/write system call when EAGAIN is returned.  This would dove tail well with the threads-per-disk work14:54
*** corvus is now known as jeblair14:59
*** B4rker has quit IRC15:00
*** Guest83839 has joined #openstack-swift15:00
*** charlesw_ has joined #openstack-swift15:02
*** charlesw has quit IRC15:03
*** charlesw_ is now known as charlesw15:03
*** minwoob has quit IRC15:12
*** B4rker has joined #openstack-swift15:12
*** david-ly_ is now known as david-lyle15:21
*** leopoldj has quit IRC15:30
*** zhill_ has joined #openstack-swift15:32
*** lyrrad has quit IRC15:33
*** foexle has quit IRC15:36
*** B4rker has quit IRC15:37
*** gyee has joined #openstack-swift15:48
*** _bluev has joined #openstack-swift15:52
*** B4rker has joined #openstack-swift15:58
*** josed has joined #openstack-swift15:59
*** _bluev has left #openstack-swift15:59
notmynamegood morning16:00
notmynameI'm helping with a swift workshop today, so i'll be in and out today16:00
*** minwoob has joined #openstack-swift16:02
*** bkopilov has quit IRC16:02
jrichlinotmyname: have fun!16:02
*** B4rker has quit IRC16:03
*** mmcardle has quit IRC16:04
*** B4rker has joined #openstack-swift16:06
*** lyrrad has joined #openstack-swift16:06
acolesmattoliverau: hi i ran across this pycharm tip recently, you may be interested http://confluence.jetbrains.com/display/IDEADEV/Inotify+Watches+Limit. i have sometimes seen it fail to keep up with me switching between branches.16:08
*** Guest83839 has quit IRC16:08
*** bkopilov has joined #openstack-swift16:08
MooingLemurclayg: might you have about 5-10 minutes today to take a glance at something I've been seeing that might be related to the fragment loss16:09
*** annegentle has joined #openstack-swift16:09
*** admin6 has quit IRC16:10
*** jistr has quit IRC16:11
*** bkopilov has quit IRC16:12
*** petertr7 is now known as petertr7_away16:12
*** jordanP has quit IRC16:14
*** annegentle has quit IRC16:14
*** petertr7_away is now known as petertr716:24
*** bkopilov has joined #openstack-swift16:31
*** petertr7 is now known as petertr7_away16:37
*** charlesw_ has joined #openstack-swift16:43
*** charlesw has quit IRC16:46
*** charlesw_ is now known as charlesw16:46
*** B4rker has quit IRC16:55
*** B4rker has joined #openstack-swift16:57
*** B4rker has quit IRC17:00
*** ptb has quit IRC17:03
*** annegentle has joined #openstack-swift17:10
*** bi_fa_fu has quit IRC17:13
*** zhill_ has quit IRC17:17
*** annegentle has quit IRC17:17
openstackgerritMerged openstack/python-swiftclient: Compare each chunk of large objects when downloading  https://review.openstack.org/17221217:18
*** annegentle has joined #openstack-swift17:21
*** mmcardle has joined #openstack-swift17:21
*** hseipp has left #openstack-swift17:25
*** josed has quit IRC17:27
openstackgerritAlistair Coles proposed openstack/swift: Tighten up overall metadata size tests  https://review.openstack.org/18726417:28
*** petertr7_away is now known as petertr717:29
acolesjrichli: ^^ curious if that passes with your backend metadata checking17:29
acolesjrichli: gotta go now, chat tomorrow17:30
*** acoles is now known as acoles_away17:30
*** mmcardle has quit IRC17:31
*** dimasot has joined #openstack-swift17:31
*** zhill_ has joined #openstack-swift17:31
*** zhill_ has quit IRC17:33
*** zhill_ has joined #openstack-swift17:33
*** zhill_ has quit IRC17:36
*** zhill_ has joined #openstack-swift17:36
*** fthiagogv has joined #openstack-swift17:41
openstackgerritThiago da Silva proposed openstack/swift-bench: add storage policy option  https://review.openstack.org/18459017:56
*** aix has quit IRC18:04
*** wer has quit IRC18:05
*** lcurtis has joined #openstack-swift18:05
pelusetdasilva, nice... FYI (not to advocate one over another but so options are understood) cosbench has that capabililty as well18:06
peluseMooingLemur, missed you the other day.  still interested in what you're finding, have company here all week though so limited time but please post whatever you've got and I'll look to18:08
tdasilvapeluse: cool! honestly I haven't used cosbench yet but would like to give it a try sometime. I just started using swift-bench recently to do some quick testing during development. I understand cosbench is a bit more robust, correct?18:08
*** geaaru has quit IRC18:08
MooingLemurpeluse: yeah, I was hoping nothing happened to you, since I didn't see you since :P18:08
lcurtishello all, trying to come up with a drive failure script aka swift-drive audit...is there a best practice whereby a drive should absolutely be unmounted for swift to properly fail over?18:09
*** lpabon has joined #openstack-swift18:09
claygMooingLemur: yeah i'm about to be on a train - so get some notes together and i'll look at it right away!18:16
*** josed has joined #openstack-swift18:18
minwoobtdasilva: I noticed that you've reviewed this patch, and requested some unit tests -- do you have some suggestions that you have in mind?18:22
minwoobtdasilva: https://review.openstack.org/#/c/184113/18:22
tdasilvaminwoob: 1 sec..looking...18:26
*** josed has quit IRC18:30
*** josed has joined #openstack-swift18:31
tdasilvaminwoob: I was thinking of something like: test.unit.obj.test_diskfile:TestECDiskFile.test_commit_raises_DiskFileErrors18:42
tdasilvaminwoob: seems to be testing the failure scenarios for the fsync, maybe could add some tests for fsync_dir too? just an idea...18:42
minwoobtdasilva: I see. Looking at it now.18:48
*** dimasot has quit IRC18:50
*** nadeem has joined #openstack-swift18:51
*** silor1 has joined #openstack-swift18:54
*** fthiagogv has quit IRC18:54
*** silor has quit IRC18:55
*** B4rker has joined #openstack-swift18:55
*** fthiagogv has joined #openstack-swift18:56
*** B4rker has quit IRC18:57
*** B4rker has joined #openstack-swift18:58
*** silor1 has quit IRC18:58
*** annegentle has quit IRC19:00
*** B4rker has quit IRC19:02
*** B4rker has joined #openstack-swift19:02
openstackgerritSamuel Merritt proposed openstack/swift: EC: support multiple ranges for GET requests  https://review.openstack.org/17349719:02
*** petertr7 is now known as petertr7_away19:05
*** petertr7_away is now known as petertr719:06
MooingLemurclayg: I'll repaste what I said on Thursday19:09
MooingLemurclayg: As it turns out, I lost a bunch of data in one container but lost nothing in another.  Both on the same storage policy.  I wonder if it was really broken all along when the ring wasn't properly dispersed for EC and that's when I uploaded those objects. that's the only difference between the two containers I can think of (since one didn't lose anything)19:10
*** tab__ has joined #openstack-swift19:10
MooingLemurclayg: here are some logs.  For some reason, even though it should be reconstructing this object, it's not.  https://bpaste.net/show/5064e2d08c3819:10
MooingLemurpeluse: ^19:11
openstackgerritSamuel Merritt proposed openstack/swift: Don't quarantine on read_metadata ENOENT  https://review.openstack.org/18711619:11
*** acampbell has joined #openstack-swift19:12
*** acampbel11 has joined #openstack-swift19:12
MooingLemurWhen I saw that log, I decided to just leave the cluster be until I could hopefully get another pair of eyes on it, because that behavior does not seem right at all.19:12
MooingLemurFYI, There are 5 storage servers, the 2nd to last digit group for the IP address for the five storage servers are 806 809 813 814 81519:14
*** zul has quit IRC19:14
*** mjseger has joined #openstack-swift19:16
jrichliacoles_away: thanks, I will check it out and let you know19:16
*** erlon has joined #openstack-swift19:18
*** doxavore has joined #openstack-swift19:22
*** mwheckmann has quit IRC19:31
mjsegerI'd like to do a little cleanup on swift-get-nodes but I can't understand what it's supposed to do.  for example the help and even the code say you can specify a container without an account but I have no idea how it would know since all the args are positional19:33
*** annegentle has joined #openstack-swift19:33
*** annegentle has quit IRC19:35
*** annegentle has joined #openstack-swift19:35
mjsegerre swift-get-nodes, rereading my post it's not that I don't understand what is does, because I do19:37
wbhuberpeluse: Regarding the patch for "log warnings when PUTs are coming in < segment size" on trello under discussions, should i put it on hold for a bit?19:37
wbhuberhttps://review.openstack.org/#/c/184325/19:38
mjsegerrather I'm not sure how one specifies various combinations of parameters and which are legal adn whcih are not.  for example, how does one call it with a container and no account?19:38
MooingLemurmjseger: that doesn't make sense to me, you need to qualify a container with the account that it resides in19:42
*** treyd_ is now known as treyd19:43
mjseger:MooingLemur I agree but there is code that says if the account is none and the container isn't then...  same thing for missing container but specified objects19:43
tdasilvamjseger: isn't this the usage help? "Usage: %prog [-a] <ring.gz> <account> [<container>] [<object>]"19:51
tdasilvai'm reading that as account is required19:51
*** annegentle has quit IRC19:51
MooingLemurI wonder if they really mean <account> [<container [<object>]]19:52
*** acampbell has quit IRC19:55
tdasilvaMooingLemur: i think so19:56
*** acampbel11 has quit IRC19:56
mjsegerMooingLemur: to elaborate a little, my original plan was to change the help from <account>[container][object] to <account>[container[object]] to preserve the positions19:56
MooingLemurthat makes complete sense to me, but there may be some piece of knowledge about it that I'm missing19:57
*** fbo has quit IRC19:59
*** B4rker has quit IRC20:02
*** fbo has joined #openstack-swift20:04
*** charlesw has quit IRC20:10
mjsegerMooingLemur: so riddle me this - it I want to run get-nodes and pass it an account/object but no container, what does the syntax even look like?20:11
*** charlesw has joined #openstack-swift20:11
*** wer has joined #openstack-swift20:11
*** proteusguy has quit IRC20:12
mjsegerother questions that come to mind are do the account/container/ring names always have to contain 'account', 'container' or 'object'?  it looks like code is in there to check, but I'd put in code to say the filename isn't recognizable as a ring name20:12
mjsegerthe man page and help don't say a word about any of this20:13
*** B4rker has joined #openstack-swift20:13
mjsegerand to make matters worse, if you run it and are missing an arg, all it does is give you the message "Attention! mismatch between ring and item detected!" which made me think my ring was corrupt20:14
*** tab__ has quit IRC20:16
*** annegentle has joined #openstack-swift20:19
pelusewbhuber, regarding that patch I'd suggest doing nothing.  The loggins is just calling attention to the fact there are objects coming in smaller than the segment size.  W/o a lot of real data I don't know that anyone can make a case that this is useful information20:20
*** zhill_ has quit IRC20:20
*** thurloat is now known as thurloat_isgone20:21
*** B4rker has quit IRC20:22
MooingLemurmjseger: I don't think that makes any sense either20:23
MooingLemurany object path needs account+container as well, and all of that is in the hash20:24
mjsegerMooingLemur: the reason I started going down this path was to just change the error message but further digging led me to looking more closely at the code.  right now I'd like to hear from someone in the know what the actual intended behavior is supposed to be20:25
mjsegerperhaps some of the checks are simply obsolete and should be removed, but as I said I just don't know enough about the intended functionality20:26
MooingLemurI'm guessing if what you're saying is true, it's code cruft.  I haven't really peeked too deep into it.20:27
MooingLemurI'm definitely not an authority :P20:27
*** B4rker has joined #openstack-swift20:34
*** HiramAbif has joined #openstack-swift20:35
*** petertr7 is now known as petertr7_away20:38
*** otoolee has quit IRC20:38
*** lpabon has quit IRC20:38
*** petertr7_away is now known as petertr720:41
*** otoolee has joined #openstack-swift20:43
openstackgerritOpenStack Proposal Bot proposed openstack/python-swiftclient: Updated from global requirements  https://review.openstack.org/8925020:51
openstackgerritOpenStack Proposal Bot proposed openstack/swift: Updated from global requirements  https://review.openstack.org/8873620:51
*** flaper87 has quit IRC21:07
*** petertr7 is now known as petertr7_away21:11
*** flaper87 has joined #openstack-swift21:12
*** flaper87 has quit IRC21:12
*** flaper87 has joined #openstack-swift21:12
*** flaper87 has quit IRC21:19
*** flaper87 has joined #openstack-swift21:20
*** erlon has quit IRC21:31
*** fthiagogv has quit IRC21:38
*** thurloat_isgone is now known as thurloat21:42
*** B4rker has quit IRC21:43
openstackgerritMinwoo Bae proposed openstack/swift: After the .durable has been written, fsync the directory.  https://review.openstack.org/18411321:44
*** marzif_ has quit IRC21:50
*** annegentle has quit IRC21:51
*** jrichli has quit IRC21:53
*** B4rker has joined #openstack-swift21:55
*** mmcardle has joined #openstack-swift21:55
*** blmartin has quit IRC21:55
*** mmcardle has quit IRC21:56
*** mmcardle has joined #openstack-swift22:03
*** openstackgerrit has quit IRC22:07
*** doxavore has quit IRC22:07
*** openstackgerrit has joined #openstack-swift22:07
*** annegentle has joined #openstack-swift22:09
*** annegentle has quit IRC22:16
*** mmcardle has quit IRC22:17
*** charlesw has quit IRC22:17
mattoliverauMorning22:17
mattoliverauacoles: cool thanks for the tip :)22:18
*** chlong has quit IRC22:18
minwoobacoles_away: Could you revisit the hash_cleanup_listdir() change, when you can? https://review.openstack.org/#/c/178317/22:19
*** B4rker has quit IRC22:19
*** mlanner has quit IRC22:21
minwoobacoles_away: Patch Set 3 was based on your recommendation; patch set 4 is based on some comments that were made after you approved patch set 3. I believe we will need a patch set 5 that looks closer to your suggestions that resulted in patch set 3.22:21
*** mlanner has joined #openstack-swift22:25
*** josed has quit IRC22:25
*** lcurtis has quit IRC22:37
*** dmorita has joined #openstack-swift22:38
*** openstackgerrit has quit IRC22:51
*** openstackgerrit has joined #openstack-swift22:51
*** wbhuber has quit IRC22:52
*** zul has joined #openstack-swift23:01
*** nadeem has quit IRC23:02
*** km has joined #openstack-swift23:05
*** agentle1 has joined #openstack-swift23:10
*** minwoob has quit IRC23:16
*** kei_yama has joined #openstack-swift23:17
*** zaitcev has joined #openstack-swift23:27
*** ChanServ sets mode: +v zaitcev23:27
*** agentle1 has quit IRC23:30
*** chlong has joined #openstack-swift23:41
*** agentle1 has joined #openstack-swift23:49
*** agentle1 has quit IRC23:49
*** charlesw has joined #openstack-swift23:51
*** ho has joined #openstack-swift23:51
*** fthiagogv has joined #openstack-swift23:55

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!