my cell server keeps crashing

27 views
Skip to first unread message

Dan Drake

unread,
Jun 28, 2012, 5:14:14 PM6/28/12
to sage-n...@googlegroups.com
Hello,

I'm finding the cell server to be pretty crashy. I have to restart it
quite often. I've attached a log of the most recent crash, but I also
see lots of MemoryError problems. (No logs for those, sadly.)

Any ideas? aleph seems to be rather reliable; any tips on how I can get
the KAIST server to work that well?

Thanks!

Dan

--
--- Dan Drake
----- http://mathsci.kaist.ac.kr/~drake
-------
cell-server.log
signature.asc

Jason Grout

unread,
Jun 28, 2012, 5:16:40 PM6/28/12
to sage-n...@googlegroups.com
On 6/28/12 4:14 PM, Dan Drake wrote:
> Hello,
>
> I'm finding the cell server to be pretty crashy. I have to restart it
> quite often. I've attached a log of the most recent crash, but I also
> see lots of MemoryError problems. (No logs for those, sadly.)
>
> Any ideas? aleph seems to be rather reliable; any tips on how I can get
> the KAIST server to work that well?
>

Use MongoDB? It seems that your problems are related to using sqlite
(lock errors, possibly memory problems...)

Aleph is pretty reliable (we do use MongoDB)...but I also have a script
that checks it every 2 minutes and restarts if it isn't able to do a
computation in a few seconds.

Thanks,

Jason

Jason Grout

unread,
Jun 28, 2012, 5:17:46 PM6/28/12
to sage-n...@googlegroups.com
On 6/28/12 4:14 PM, Dan Drake wrote:
> Hello,
>
> I'm finding the cell server to be pretty crashy. I have to restart it
> quite often. I've attached a log of the most recent crash, but I also
> see lots of MemoryError problems. (No logs for those, sadly.)
>
> Any ideas? aleph seems to be rather reliable; any tips on how I can get
> the KAIST server to work that well?

I guess another tip: wait for the next release, which doesn't use the
database *nearly* as much as the current version does. See the
contrib/ipython-testing directory in the master branch (and the many
pull requests are now reviewing...)

Thanks,

Jason


Reply all
Reply to author
Forward
0 new messages