That's true, I can't either. But I was able to reproduce it consistently for several days using the same method. Refresh a cached page, clear the page cache, repeat. Doing this about 3 or 4 times pushed mongrel to about 160MB foot print. Mongrel debug logging showed a massive number of 'String' and 'Array' objects never being garbage collected. The system was basically setup like this:
Redhat EL3 Apache 2.0.46 (I think) Mongrel 0.3.13 sqlite3 - ruby bindings radiant 0.5.0 Keeping everything the same and switching to postgres 8 solved the problem and it's been running stable for about 2 weeks now. So what the exact reason was, I don't know, but I do know that the sqlite3 ruby bindings were the difference between leaking and not leaking. Josh John W. Long wrote: > Josh Ferguson wrote: > >> I had bad memory leak problems using mongrel, apache, radiant and >> mongrel. I tracked it for more than a week and could reproduce it. >> Switching to postgres solved the memory leak completely. I'm not sure >> what the problem was but I would suggest staying away from the sqlite3 >> bindings as for whatever reason radiant's caching causes memory leaks >> with it. >> > > You say Radiant's caching caused memory leaks. Do you say this because > you turned it off and the problem went away or was this just a suspicion? > > I can't see any reason why the choice of database drivers and the > caching mechanism would have anything to do with each other. > > -- > John Long > http://wiseheartdesign.com > _______________________________________________ > Radiant mailing list > Radiant@lists.radiantcms.org > http://lists.radiantcms.org/mailman/listinfo/radiant > > _______________________________________________ Radiant mailing list Radiant@lists.radiantcms.org http://lists.radiantcms.org/mailman/listinfo/radiant