Displaying #cassandra-dev/2016-11-11.log:

Fri Nov 11 00:05:42 2016  kohlisankalp:Joined the channel
Fri Nov 11 00:09:53 2016  clohfink:Joined the channel
Fri Nov 11 00:14:30 2016  kohlisankalp:Joined the channel
Fri Nov 11 00:17:56 2016  zznate:its a feature to make people think carefully about their data model upfront
Fri Nov 11 00:18:42 2016  zznate:(afternoon friday troll just for krut)
Fri Nov 11 00:18:54 2016  zznate:srsly, that is terrible and bad.
Fri Nov 11 00:21:16 2016  zznate:also sounds somewhat related to CASSANDRA-11050 we found test driving 3.0 last year
Fri Nov 11 00:21:16 2016  CassBotJr:https://issues.apache.org/jira/browse/CASSANDRA-11050 (Resolved; Fixed; 3.0.4, 3.4): "SSTables not loaded after dropping column"
Fri Nov 11 00:28:04 2016  gila:Joined the channel
Fri Nov 11 00:34:31 2016  mstepura:Joined the channel
Fri Nov 11 00:48:10 2016  gila_:Joined the channel
Fri Nov 11 00:55:15 2016  aaron_morton:Joined the channel
Fri Nov 11 01:30:10 2016  kohlisankalp:Joined the channel
Fri Nov 11 01:33:12 2016  krut:does sound similar to that. commitlog was just one of the issues they had.
Fri Nov 11 01:55:21 2016  kohlisankalp:Joined the channel
Fri Nov 11 02:15:21 2016  aaron_morton:Joined the channel
Fri Nov 11 02:19:41 2016  clohfink:Joined the channel
Fri Nov 11 02:21:01 2016  clohfink:Joined the channel
Fri Nov 11 02:22:53 2016  clohfink:Joined the channel
Fri Nov 11 02:25:16 2016  clohfink_:Joined the channel
Fri Nov 11 02:32:04 2016  mstepura:Joined the channel
Fri Nov 11 02:33:00 2016  dikang:Joined the channel
Fri Nov 11 03:52:40 2016  nickmbailey:Joined the channel
Fri Nov 11 03:58:08 2016  aaron_morton:Joined the channel
Fri Nov 11 04:28:04 2016  Vijay:Joined the channel
Fri Nov 11 04:47:08 2016  nickmbailey:Joined the channel
Fri Nov 11 04:48:14 2016  JayZhuang:Joined the channel
Fri Nov 11 05:27:03 2016  dikang:Joined the channel
Fri Nov 11 05:32:51 2016  kohlisankalp:Joined the channel
Fri Nov 11 05:35:16 2016  dikang:Joined the channel
Fri Nov 11 05:42:28 2016  dikang:Joined the channel
Fri Nov 11 06:18:42 2016  dikang:Joined the channel
Fri Nov 11 06:34:13 2016  dikang_:Joined the channel
Fri Nov 11 06:35:16 2016  nickmbailey:Joined the channel
Fri Nov 11 06:52:09 2016  mstepura:Joined the channel
Fri Nov 11 06:52:52 2016  vince_:Joined the channel
Fri Nov 11 07:29:37 2016  nickmbailey:Joined the channel
Fri Nov 11 07:33:23 2016  minimarcel__:Joined the channel
Fri Nov 11 07:34:57 2016  kvaster_:Joined the channel
Fri Nov 11 07:36:25 2016  jfarrell:Joined the channel
Fri Nov 11 07:47:31 2016  kvaster:Joined the channel
Fri Nov 11 08:14:33 2016  spodkowinski:Joined the channel
Fri Nov 11 08:23:37 2016  nickmbailey:Joined the channel
Fri Nov 11 08:32:10 2016  spodkowinski:Joined the channel
Fri Nov 11 08:39:17 2016  iamaleksey:krut: it's new to 3.0; until then only affected dropped collection column (and recreating those was forbidden)
Fri Nov 11 08:46:41 2016  zaller:Joined the channel
Fri Nov 11 09:17:36 2016  nickmbailey:Joined the channel
Fri Nov 11 09:19:04 2016  minimarcel__:Joined the channel
Fri Nov 11 09:21:13 2016  clohfink:Joined the channel
Fri Nov 11 10:04:56 2016  gila:Joined the channel
Fri Nov 11 11:49:24 2016  vince_:Joined the channel
Fri Nov 11 12:00:04 2016  nickmbailey:Joined the channel
Fri Nov 11 12:36:54 2016  kvaster:Joined the channel
Fri Nov 11 14:55:44 2016  adamholmberg:Joined the channel
Fri Nov 11 15:20:22 2016  minimarcel__:Joined the channel
Fri Nov 11 15:22:23 2016  ArrgntBstrd:Joined the channel
Fri Nov 11 15:25:16 2016  clohfink:Joined the channel
Fri Nov 11 15:41:51 2016  Vijay:Joined the channel
Fri Nov 11 15:51:34 2016  Buelent:Joined the channel
Fri Nov 11 15:53:05 2016  Buelent:Left the channel
Fri Nov 11 16:01:59 2016  clohfink:Joined the channel
Fri Nov 11 16:07:17 2016  Vijay:Joined the channel
Fri Nov 11 16:14:41 2016  exlt:thought I started those votes Monday, so 3.10 close is a few hours early, but I don't think the result will change.
Fri Nov 11 16:15:28 2016  exlt:3.0.10 vote will stay open, since jake just commented
Fri Nov 11 16:17:16 2016  exlt:although that reopened bug is pretty old and probably an edge case - I'm not really sure about how critical it really is - sway me :)
Fri Nov 11 16:17:41 2016  exlt:not saying it shouldn't be fixed, but it's certainly not new since 3.0.9
Fri Nov 11 16:21:05 2016  exlt:also, I don't have a problem rerolling 3.0.10 after #11039 gets committed - is that something you're working on today, tjake?
Fri Nov 11 16:21:05 2016  CassBotJr:https://issues.apache.org/jira/browse/CASSANDRA-11039 (Reopened; Unresolved; 3.0.x): "SegFault in Cassandra"
Fri Nov 11 16:23:58 2016  tjake:yep
Fri Nov 11 16:24:16 2016  exlt:thanks
Fri Nov 11 16:26:15 2016  aaron_morton:Joined the channel
Fri Nov 11 16:26:58 2016  aaron_morton:Joined the channel
Fri Nov 11 16:45:28 2016  gila:Joined the channel
Fri Nov 11 16:50:37 2016  tjake:made you reviewer just checking the dtests dont set offheap_buffers
Fri Nov 11 16:53:03 2016  ArrgntBstrd:Joined the channel
Fri Nov 11 16:54:21 2016  vjarnot:Joined the channel
Fri Nov 11 16:55:44 2016  exlt:git grep says no - thanks for the changes/news edits there, too :)
Fri Nov 11 16:56:41 2016  beobal:exlt: could you hold the re-roll for a little while longer? I need to ninja in a tiny fix to the #12813 patch I committed earlier, just double checking tests now
Fri Nov 11 16:56:43 2016  CassBotJr:https://issues.apache.org/jira/browse/CASSANDRA-12813 (Reopened; Unresolved; 2.2.9, 3.10, 3.11): "NPE in auth for bootstrapping node"
Fri Nov 11 16:57:01 2016  exlt:yep yep!
Fri Nov 11 16:57:06 2016  beobal:thanks
Fri Nov 11 17:00:50 2016  Vijay:Joined the channel
Fri Nov 11 17:01:33 2016  exlt:72 hours is officially up in just over 3 hours - I could smack the vote with a hammer, or there could also be other -1 votes :)
Fri Nov 11 17:04:38 2016  tjake:Any binding -1 kills the vote
Fri Nov 11 17:05:01 2016  tjake:" -1 votes are vetos and kill the proposal dead until all vetoers withdraw their -1 votes."
Fri Nov 11 17:05:10 2016  tjake:http://apache.org/foundation/voting.html
Fri Nov 11 17:05:48 2016  ptnapoleon: That's for patches
Fri Nov 11 17:05:51 2016  ptnapoleon:It's different for releases
Fri Nov 11 17:05:53 2016  ptnapoleon:"Votes on whether a package is ready to be released use majority approval"
Fri Nov 11 17:05:55 2016  ptnapoleon:Same link
Fri Nov 11 17:06:41 2016  tjake:ah
Fri Nov 11 17:06:44 2016  tjake:TIL
Fri Nov 11 17:07:02 2016  ptnapoleon:Yeah, I thought the same thing until ay corrected me
Fri Nov 11 17:07:24 2016  exlt:yeah, I learned this a couple months ago :)
Fri Nov 11 17:07:46 2016  tjake:I wish I knew this when I was release manager ;)
Fri Nov 11 17:07:53 2016  exlt:heh
Fri Nov 11 17:07:54 2016  tjake:so much power
Fri Nov 11 17:08:22 2016  iamaleksey:@exlt: need an extra -1? I don't think I voted on that one
Fri Nov 11 17:08:45 2016  exlt:haven't counted, but there are definitely more +1 votes
Fri Nov 11 17:09:08 2016  exlt:looks like 3-2 now
Fri Nov 11 17:09:35 2016  mebigfatguy:that seems backwards
Fri Nov 11 17:09:52 2016  mebigfatguy:why would it be harder to release a patch, then a release?
Fri Nov 11 17:10:13 2016  exlt:I understand that to mean like +1 on a JIRA ticket
Fri Nov 11 17:10:41 2016  ptnapoleon:I assume because a release is made up of patches that weren't vetoed, and thus shouldn't be too controversial?
Fri Nov 11 17:10:52 2016  ptnapoleon:Releases are easier to re-roll than patches are to revert?
Fri Nov 11 17:11:29 2016  iamaleksey:mebigfatguy: it is backwards. What kind of direction would you expect from ASF rules, really? (:
Fri Nov 11 17:12:38 2016  mebigfatguy:heh
Fri Nov 11 17:13:26 2016  exlt:is my understanding maybe limited, thinking of this as just a patch? does "code modification" mean something like a feature? ie. vnodes vs. staying with single-token
Fri Nov 11 17:15:06 2016  ostefano:Joined the channel
Fri Nov 11 17:15:08 2016  mstepura:Joined the channel
Fri Nov 11 17:17:05 2016  exlt:the lazy consensus exmple at the end uses a patch example
Fri Nov 11 17:24:29 2016  nickmbailey:Joined the channel
Fri Nov 11 17:35:44 2016  dikang:Joined the channel
Fri Nov 11 17:36:00 2016  minimarcel__:Joined the channel
Fri Nov 11 17:39:17 2016  EdwardCapriolo:Joined the channel
Fri Nov 11 17:40:13 2016  mstepura:Joined the channel
Fri Nov 11 17:45:44 2016  JayZhuang:Joined the channel
Fri Nov 11 17:47:39 2016  ptnapoleon:zznate: Should this have also gone into 3.0? https://github.com/apache/cassandra/commit/456c0280e90c92724669c4aba6ae9461af02d663
Fri Nov 11 17:51:16 2016  beobal:exlt: that ninja fix for 12813 is committed now
Fri Nov 11 17:51:30 2016  ptnapoleon:The dtest is failing on 3.0, because that only went into 3.x and trunk, and i'm wondering if i should just change the tests, or wait for that ninja to be pushed into 3.0 as well
Fri Nov 11 17:53:42 2016  nickdevp:Joined the channel
Fri Nov 11 18:01:50 2016  nickdevp:Left the channel
Fri Nov 11 18:18:32 2016  nickmbailey:Joined the channel
Fri Nov 11 18:28:19 2016  zaller:Joined the channel
Fri Nov 11 18:41:47 2016  aaron_morton:Joined the channel
Fri Nov 11 18:54:20 2016  kohlisankalp:Joined the channel
Fri Nov 11 20:01:47 2016  Vijay_:Joined the channel
Fri Nov 11 20:04:50 2016  zznate:ptnapolean omg, yes
Fri Nov 11 20:04:53 2016  zznate:dammit
Fri Nov 11 20:05:21 2016  zznate:no, def ninja fix is going to ugly error everytime that logger method is called
Fri Nov 11 20:05:45 2016  ptnapoleon:I do not understand that third message
Fri Nov 11 20:05:53 2016  zznate:haha. its saturday morning here.
Fri Nov 11 20:06:06 2016  zznate:yes pls ninja fix that
Fri Nov 11 20:06:18 2016  ptnapoleon:Oh right, it is saturday
Fri Nov 11 20:06:20 2016  ptnapoleon:thanks!
Fri Nov 11 20:06:20 2016  ptnapoleon:sorry
Fri Nov 11 20:06:34 2016  zznate:thank u!! good catch, that would have sucked.
Fri Nov 11 20:08:09 2016  ptnapoleon:Can someone with commit access make that fix? ^
Fri Nov 11 20:08:17 2016  ptnapoleon:It's saturday for nate, and I am not a committer
Fri Nov 11 20:12:39 2016  zznate:ptnapoleon .. actually do you have an example of the stack trace that threw??
Fri Nov 11 20:12:53 2016  ptnapoleon:Oh it's just failing the test, as mentioned in the ticket
Fri Nov 11 20:12:58 2016  ptnapoleon:Because the log output changed
Fri Nov 11 20:13:12 2016  zznate:oh.
Fri Nov 11 20:13:30 2016  zznate:i thought i missed a parameter on the slf4j replacement
Fri Nov 11 20:20:32 2016  mstepura:Joined the channel
Fri Nov 11 20:36:39 2016  minimarcel__:Joined the channel
Fri Nov 11 21:00:20 2016  kohlisankalp:Joined the channel
Fri Nov 11 21:21:46 2016  jfarrell:Joined the channel
Fri Nov 11 21:43:48 2016  kohlisankalp:Joined the channel
Fri Nov 11 21:48:48 2016  zaller1:Joined the channel
Fri Nov 11 21:55:25 2016  kohlisankalp:Joined the channel
Fri Nov 11 22:02:26 2016  vjarnot:Joined the channel
Fri Nov 11 22:05:59 2016  JayZhuang:Joined the channel
Fri Nov 11 22:11:23 2016  zznate:^i've been thinking about this while drinking coffee this AM... (since it's a nice leasurely saturday morning here)
Fri Nov 11 22:11:35 2016  zznate:adhoc parsing of log statements for test verification is ghetto.
Fri Nov 11 22:13:03 2016  zznate:primarily 'cause it limits flexibility of checking/verification to something that can only be verified on the filesystem
Fri Nov 11 22:14:50 2016  zznate:in some of the contexts where we are doing this in dtests, we could add (potentially check an already existing) metric.
Fri Nov 11 22:39:09 2016  exlt:that's a thought that has come up before, but if I recall, I think that requires dtest to get some porting to jython or to consume json through jolokia
Fri Nov 11 22:40:41 2016  exlt:I also recall race conditions here and there where JMX says something is ready, but it's not really ready yet
Fri Nov 11 22:41:09 2016  exlt:even in logs, this has been true
Fri Nov 11 22:41:24 2016  exlt:so maybe that's not a decent bullet point
Fri Nov 11 22:42:53 2016  zaller:Joined the channel
Fri Nov 11 22:49:42 2016  aaron_morton:Joined the channel
Fri Nov 11 22:57:03 2016  kohlisankalp:Joined the channel
Fri Nov 11 22:58:06 2016  JayZhuang:Joined the channel
Fri Nov 11 23:03:31 2016  zznate:krut, did you create an issue for that 'drop and re-add column with new type = sad cluster' thing ?
Fri Nov 11 23:04:02 2016  dikang:Joined the channel
Fri Nov 11 23:09:53 2016  kohlisan_:Joined the channel
Fri Nov 11 23:17:10 2016  aaron_morton:Joined the channel
Fri Nov 11 23:23:15 2016  mal``:Joined the channel
Fri Nov 11 23:36:50 2016  jeffj:pretty sure it already exists
Fri Nov 11 23:43:27 2016  nickmbailey:Joined the channel

Comments