Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2015-04-18

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:12 kdhananjay joined #gluster-dev
01:51 nishanth joined #gluster-dev
02:44 atinmu joined #gluster-dev
02:57 gem joined #gluster-dev
03:42 soumya joined #gluster-dev
06:06 lalatenduM joined #gluster-dev
06:32 atinmu joined #gluster-dev
06:45 lalatenduM joined #gluster-dev
07:17 gem joined #gluster-dev
09:01 soumya joined #gluster-dev
10:34 gem joined #gluster-dev
10:43 gem joined #gluster-dev
10:45 atinmu joined #gluster-dev
10:48 pcaruana joined #gluster-dev
10:55 atinmu joined #gluster-dev
11:11 dlambrig_ joined #gluster-dev
11:11 dlambrig_ v
11:11 dlambrig_ left #gluster-dev
12:26 foster joined #gluster-dev
14:20 gem joined #gluster-dev
14:48 gem joined #gluster-dev
14:53 gem joined #gluster-dev
15:04 vipulnayyar joined #gluster-dev
16:00 kshlm joined #gluster-dev
16:28 wushudoin joined #gluster-dev
16:48 gem joined #gluster-dev
17:43 rafi joined #gluster-dev
18:09 * JustinCl1ft headbutts desk
19:18 soumya joined #gluster-dev
19:59 misc JustinCl1ft: why ?
21:13 JustinCl1ft Gerrit ;)
21:13 JustinCl1ft misc: Well, in this instance H2.
21:14 JustinCl1ft The built in "not recommended for production" database that it's using
21:14 JustinCl1ft We have some bad data "somewhere" (!) in the database, which is making things hard
21:14 JustinCl1ft And the tools for accessing H2 are um... lousy
21:15 JustinCl1ft So, I'm writing H2 -> PostgreSQL conversion stuff atm
21:15 JustinCl1ft PL/PgSQL code right now (not complex variety fortunately)
21:16 misc bad data, being defined by ?
21:18 JustinCl1ft Upgrading past version 2.8.6.1 means we _have_ to define a secondary index on the database, then run the "reindex" command.
21:18 JustinCl1ft This spits out an error on the way through, though it does complete
21:19 JustinCl1ft The error is about it not being able to find a "glusterfs/" repo.
21:20 JustinCl1ft So, there's a bad <something> in the data to do with "glusterfs/".  We of course shows up in a few places (60k+)... so it's a bit hard to narrow down.
21:21 JustinCl1ft When doing this upgrade, my admin capabilities change afterwards.  I can no longer merge changes into our main repo.
21:21 JustinCl1ft I can't tell if it's related to the error message or not. :/.
21:22 JustinCl1ft ... and a bunch of other side problems as well.  Which all pretty much boil down to -> We should migrate to PostgreSQL if at all possible. ;)
21:23 JustinCl1ft ... and I'm still working on making that possible.  If I get it working tonight with the older database backup, then it should work ok tomorrow with the real production one.
21:24 JustinCl1ft If I don't get it working tonight, then tomorrow I'll just upgrade the production one in place using H2, and we'll have to hope for the best + fix anything screwy as it comes along afterwards. :(
21:24 JustinCl1ft I suspect we'll be creating some scripting or something to validate the database data in the coming days/weeks.
21:25 JustinCl1ft That's going to take a bunch of effort though :(
21:27 JustinCl1ft Actually, what I'll probably do is just upgrade us to version 2.8.6.1 (the one before we have to use a secondary index), as that has a GitHub OAuth plugin available.
21:28 JustinCl1ft That would work for us for now, and give us breathing space to figure out the deeper database problem.
21:28 JustinCl1ft It just depends on what works and what goes wrong tomorrow with the real database. ;)
21:32 JustinCl1ft Hmmm...
21:33 JustinCl1ft misc: I just realised the version of PostgreSQL on review.gluster.org is very old.  8.1.
21:33 JustinCl1ft Gerrit doesn't list a minimum PG version... so hopefully that's ok.
21:40 misc JustinCl1ft: I do not imagine gerrit using advanced feature of pgsql
21:46 JustinCl1ft Yeah
21:46 JustinCl1ft Gah.  My PL/pgSQL knowledge is so rusty, and I'm tired. ;)
21:49 JustinCl1ft I might write it in PL/Python instead
21:50 JustinCl1ft Prob better not to.  PL/pgSQL is more standard.
22:32 JustinCl1ft Cool.  Got the PL/pgSQL function working. (it's been years ;>)
22:39 JustinCl1ft https://gist.github.com/justinclift/42dad0372e1aaa4f366c
22:44 JustinCl1ft Muaaa Haaa Haaa
22:45 JustinCl1ft H2 data seems to be importing ok using the old backup database
22:45 JustinCl1ft And that PL/pgSQL function is working too :)
22:48 JustinCl1ft Gah
22:48 JustinCl1ft I'll need to write another conversion function
22:49 JustinCl1ft Still, it's close to workable as-is database conversion wise
22:49 JustinCl1ft Just need one to properly convert \n and similar to actual newlines
22:49 JustinCl1ft So the import of newlines works
23:01 JustinCl1ft Trying out the Gerrit upgrade procedure using PostgreSQL instead of H2 database now, to see if there's any difference
23:04 misc at least, with pg, you can investigate more easily
23:06 JustinCl1ft yeah
23:07 JustinCl1ft There is existing Gerrit code around for merging users too (for PostgreSQL setups)
23:07 JustinCl1ft Which we'll likely need
23:36 wushudoin joined #gluster-dev
23:42 JustinCl1ft Muaaa Haaa.  Found the fixed the invalid data lines.
23:42 JustinCl1ft There were 5.
23:42 JustinCl1ft Reindexed again, no errors showing p.
23:42 JustinCl1ft up
23:42 JustinCl1ft Now to see if it works any better.
23:42 * JustinCl1ft hopes
23:46 JustinCl1ft Interestingly, the bad CR's can now be viewed (after fixing them)... and they all turn out to be from the same developer
23:46 JustinCl1ft "shylesh kumar"
23:50 misc CR ?
23:51 JustinCl1ft Change Request
23:52 misc mhhh
23:52 JustinCl1ft Gerrit terminology for a GitHub PR
23:52 misc curious, could be some db corruption on his account ?
23:52 JustinCl1ft Doesn't seem to be
23:53 JustinCl1ft I'm guessing someone edited the database directly or maybe there was a bug in Gerrit or something
23:53 * JustinCl1ft shrugs
23:53 JustinCl1ft No idea :)
23:57 JustinCl1ft k, that's enough from me for tonight.  I need to be functional tomorrow, so need to hit the sack. :)
23:57 JustinCl1ft 'nite dude
23:58 misc 'nit

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary