On Tue, 26 Aug 2014, Joachim Breitner wrote: > Hi, > > > Am Dienstag, den 26.08.2014, 16:09 -0700 schrieb Alexander Wirt: > > On Tue, 26 Aug 2014, Joachim Breitner wrote: > > > Am Sonntag, den 24.08.2014, 21:16 -0700 schrieb Alexander Wirt: > > > > I don't exactly know what fetchdata is, but it kills alioth. So you will > > > > either: > > > > > > > > a) rewrite it to be not to ressource intensive > > > > b) runs the process with lower priority > > > > c) run it somewhere else > > > > > > it is part of PET, the package tracker, that we run for the Debian > > > Haskell Group, and I don’t know much about it. > > > > > > @PET-Devel: Can you help me figure out what’s going here? > > > > > > (Is PET even actively maintained these days?) > > And does this script really run every hour? > > yes, and it has done that unmodified since many years. Usually, it is > fast... > > Since sunday night it said "Can't read cache:" (without more). I then > deleted the cache yesterday afternoon, which made it stop giving this > error. But from what you say it is now causing other problems? We had a hard night were alioth had a load of a few hundred. And the script was on the toplist of software eating our disk io. Alex
Attachment:
pgplXmjvUJL11.pgp
Description: PGP signature