Op Di, 2009-06-23 om 19:18 +0530 skryf Sayamindu Dasgupta: > Hello, > We at Sugarlabs (translate.sugarlabs.org) are facing a somewhat queer > issue while dealing with large files. When a multiple translators >= 3 > are working on a large PO file (etoys: with around 4370 strings), > sometimes the server becomes non-responsive for all subsequent > requests, and the only solution is to restart the server (PootleServer > manually). I'm not sure what is happening - we have Xapian indexing > enabled.. is it possible that it may be responsible somehow. I just > remove python-xapian from the server to see how it goes, and will > report back if there are no more occurrences of this type. > I did strace on a non-responsive server, and it seems to have gone on > a loop of alternating mmap() and brk() calls. > Any ideas/pointers on what may be wrong. > > Details on the server are at http://translate.sugarlabs.org/about.html
Hallo Sayamindu We started testing some ideas in IRC, but we will need some more information. If you can get a few lines of output of vmstat 1 that might help. When Translate Toolkit 1.4.0 is out, it might still be compatible with Pootle 1.2.1 - that might be another possibility. I have a few other ideas, but it would be best to try to understand the problem better. Keep well Friedel -- Recently on my blog: http://translate.org.za/blogs/friedel/en/content/pseudolocalisation-podebug-1 ------------------------------------------------------------------------------ _______________________________________________ Translate-pootle mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/translate-pootle
