zas: I have a lot of problems in python with SNI now that most of our stuff is on HTTPS
2016-04-26 11752, 2016
Gentlecat
maybe we should update at least to 2.7.9
2016-04-26 11703, 2016
UmkaDK has quit
2016-04-26 11708, 2016
Gentlecat
or even to .11, if possible
2016-04-26 11735, 2016
zas
Gentlecat: can you create a MBH ticket about this ?
2016-04-26 11744, 2016
Gentlecat
yeah, for sure
2016-04-26 11704, 2016
zas
It also means we may require 2.7.9 for picard and other python stuff
2016-04-26 11711, 2016
Gentlecat
I also had some more questions about uwsgi, but don't know how much you know
2016-04-26 11719, 2016
zas
not much
2016-04-26 11729, 2016
Gentlecat
ok, I'll do some research then
2016-04-26 11746, 2016
Gentlecat would really like to deploy CB with py3, even for testig
2016-04-26 11748, 2016
zas
but i'm interested in your findings
2016-04-26 11752, 2016
Freso
zas: I think we already switched minimum version of Picard to 2.7 a while back, so moving 0.0.9 versions up would likely not be that invasive of a change.
2016-04-26 11756, 2016
UmkaDK joined the channel
2016-04-26 11704, 2016
zas
Freso: yup
2016-04-26 11707, 2016
Gentlecat
I think zas is talking about websites, no?
2016-04-26 11708, 2016
Freso
(Or maybe I'm thinking of another Python project.)
2016-04-26 11723, 2016
Gentlecat
ones that are on sakura
2016-04-26 11725, 2016
zas
Gentlecat: nope about picard app
2016-04-26 11725, 2016
Freso
Gentlecat: I was replying to zas | It also means we may require 2.7.9 for picard and other python stuff
2016-04-26 11727, 2016
Freso
:)
2016-04-26 11728, 2016
Gentlecat
oh
2016-04-26 11754, 2016
zas
45 minutes to regain 6% of disk space on spike...
2016-04-26 11721, 2016
zas
software raid is slow
2016-04-26 11700, 2016
alastairp
:(
2016-04-26 11720, 2016
zas
this is why /tmp on its own partition is much better... too bad.
2016-04-26 11709, 2016
Freso
zas: I once said that to someone admin'ing a system, and he said he didn't want to do it because that's not how the sys admin before him that was "omg so pro" had set it up.
2016-04-26 11731, 2016
zas
Freso: from my experience, having logs and tmp on separated partitions is a must, too many bad things happen else
2016-04-26 11713, 2016
Freso
I don't disagree.
2016-04-26 11723, 2016
Freso
Au contraire, I very much agree.
2016-04-26 11732, 2016
zas
in this case, a /tmp partition would have been very fast to clean up (unmount/format/remount)
2016-04-26 11718, 2016
zas
plus big partitions are usually slower
2016-04-26 11704, 2016
zas
at this rate, we will recover whole spike's disk space in ... 11 hours. Though i expect a speed up after a while.
2016-04-26 11710, 2016
Slurpee has quit
2016-04-26 11744, 2016
Freso
Setting up /tmp as tmpfs is one of the first things I do when setting up a new machine (which makes a clean up = umount/remount).
2016-04-26 11701, 2016
alastairp
zas, load doesn't seem to be too high. Can we consider restarting uwsgi?
2016-04-26 11716, 2016
zas
yes
2016-04-26 11721, 2016
zas
done
2016-04-26 11724, 2016
alastairp
thanks
2016-04-26 11731, 2016
alastairp
I'm still confused about validate.py
2016-04-26 11736, 2016
alastairp
I can't remember what it was
2016-04-26 11714, 2016
alastairp
ohhhh. evaluate.py
2016-04-26 11729, 2016
alastairp
yes, it's running in my screen :)
2016-04-26 11706, 2016
Slurpee joined the channel
2016-04-26 11707, 2016
diana_olhovyk has quit
2016-04-26 11730, 2016
diana_olhovyk joined the channel
2016-04-26 11754, 2016
Lingo__ has quit
2016-04-26 11728, 2016
Lingo__ joined the channel
2016-04-26 11705, 2016
Lingo__ has quit
2016-04-26 11740, 2016
Lingo__ joined the channel
2016-04-26 11743, 2016
CatCat has quit
2016-04-26 11724, 2016
CatCat joined the channel
2016-04-26 11724, 2016
CatCat has quit
2016-04-26 11724, 2016
CatCat joined the channel
2016-04-26 11739, 2016
diana_olhovyk has quit
2016-04-26 11740, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11740, 2016
ZaphodBeeblebrox has quit
2016-04-26 11740, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11702, 2016
ZaphodBeeblebrox has quit
2016-04-26 11716, 2016
UmkaDK
Guys, a bit of rundown question but is there something like a maintenance script anywhere that would clean up database of all stale records?
2016-04-26 11731, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11731, 2016
ZaphodBeeblebrox has quit
2016-04-26 11731, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11724, 2016
CatQuest has quit
2016-04-26 11700, 2016
Freso
Stale records?
2016-04-26 11731, 2016
ZaphodBeeblebrox has quit
2016-04-26 11746, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11746, 2016
ZaphodBeeblebrox has quit
2016-04-26 11746, 2016
ZaphodBeeblebrox joined the channel
2016-04-26 11728, 2016
UmkaDK
The db on my sandbox has been growing steadily...
2016-04-26 11744, 2016
UmkaDK
And today it filled up its partition :(
2016-04-26 11746, 2016
ZaphodBeeblebrox has quit
2016-04-26 11739, 2016
UmkaDK
I was hoping that there is something I could do to shrink it back down a bit... otherwise I have to resize the partition, which is a bit of a bother.
2016-04-26 11742, 2016
Freso
MB db?
2016-04-26 11708, 2016
regagain joined the channel
2016-04-26 11749, 2016
JesseW joined the channel
2016-04-26 11712, 2016
UmkaDK
Yep, MB db
2016-04-26 11709, 2016
mihaitish has quit
2016-04-26 11754, 2016
yeeeargh joined the channel
2016-04-26 11739, 2016
sampsyo has quit
2016-04-26 11707, 2016
Lingo__ has quit
2016-04-26 11741, 2016
Lingo__ joined the channel
2016-04-26 11703, 2016
Freso
VACUUM it?
2016-04-26 11714, 2016
JesseW has quit
2016-04-26 11713, 2016
sampsyo joined the channel
2016-04-26 11716, 2016
MBJenkins has quit
2016-04-26 11714, 2016
kanha has quit
2016-04-26 11754, 2016
MBJenkins joined the channel
2016-04-26 11735, 2016
CatQuest joined the channel
2016-04-26 11705, 2016
CatQuest has quit
2016-04-26 11754, 2016
_Dave_ has quit
2016-04-26 11700, 2016
CatQuest joined the channel
2016-04-26 11707, 2016
Lingo__ has quit
2016-04-26 11742, 2016
Lingo__ joined the channel
2016-04-26 11720, 2016
CatQuest has quit
2016-04-26 11708, 2016
chirlu
UmkaDK: You can try VACUUM FULL, but it will take time and temporarily need more storage (because it will rewrite the tables).
2016-04-26 11728, 2016
chirlu
What are you doing on that system to create that many dead rows?
2016-04-26 11746, 2016
CatQuest joined the channel
2016-04-26 11757, 2016
kepstin
well, if it's dead rows, autovacuum should have dealt with it
2016-04-26 11711, 2016
kepstin
it must be actually growing in data - are you loading edit data or something?
2016-04-26 11735, 2016
CatQuest has quit
2016-04-26 11701, 2016
CatQuest joined the channel
2016-04-26 11701, 2016
CatQuest has quit
2016-04-26 11701, 2016
CatQuest joined the channel
2016-04-26 11750, 2016
CatQuest has quit
2016-04-26 11749, 2016
CatQuest joined the channel
2016-04-26 11701, 2016
Mineo joined the channel
2016-04-26 11709, 2016
CatQuest has quit
2016-04-26 11707, 2016
CatQuest joined the channel
2016-04-26 11735, 2016
CatQuest has quit
2016-04-26 11707, 2016
chirlu
kepstin: Not necessarily, if the old rows were still needed (long-running transaction).
2016-04-26 11745, 2016
chirlu
Another possibility would be if you simply use a *lot* of temporary rows once, since the table will not normally shrink.
