Hello list,
I am in big trouble. The spec:
- Centos 6.3 fully updated on 6 core CPU (X5675 @ 3.07GHz) and 12GB RAM
(xen vm)
- Owncloud 5.0 with mysql backend
- 250k files (40 users)
- 1 user has 113350 files
I'm updating directly from the tar.bz2 file on the server with rsync.
All my update experiences have been problematic since 4.5 (like always
forced to put maintenance mode to false manually), but at the end of the
day it was working. But this time it's a complete failure.
First I tried to go from 5.0.0 straight to 5.0.3. My mysqld process went
on 100% on a CPU for 24 hours. While "stracing" the pid I can see a lot
of this:
read(99, "\3\0\236\2\fLotsOfFilesUserh/c", 20) = 20
pread(99, "lientsync/Folder/DocumentC"..., 142, 11140924) = 142
lseek(99, 11141068, SEEK_SET) = 11141068
LotsOfFilesUser is the user with 113350 files.
Nothing in owncloud.log, neither in apache's logs :(
So I have killed the process this morning, went back to a working state
with backup from 5.0 and tried to upgrade to 5.0.1 then 5.0.2 and *bam*
same problem with 5.0.2 :(
What should I do ? Thanks for any pointer as I dont know if letting this
process running for the whole weekend will do anything good or if it is
normal that updating owncloud can take several days :(
Should I remove the user with 113k files ? If yes how (without
interfering with others users) ?
Thanks,
kfx
_______________________________________________
Owncloud mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/owncloud
