> No. Can you post the last few lines from your errors and access logs?Well, part of the problem is that this server gets so hammered - 99.9% CPU devoted to nslapd all the time, that the access logs roll over so quickly, I don''t have them any more. According to the FDSgraph charts (which are messed up because I think it''s so busy, the graphs are just a patchwork of points which aren''t joined up), it does about 70k ops per minutes, peaking at about 140k. This suddenly shot up a few weeks ago and nobody seems to know why so I''m assuming that the load is the problem. I was thinking about increasing the mem cache for FDS - worth trying? It seems to be at the default (10 Meg?) and this box has easily 3Gb doing nothing at any one time. PK
Richard Megginson
2007-Apr-05 02:33 UTC
Re: [Fedora-directory-users] Re: Another segfault ...
Philip Kime wrote:>> No. Can you post the last few lines from your errors and access logs? >> > > Well, part of the problem is that this server gets so hammered - 99.9% > CPU devoted to nslapd all the time, that the access logs roll over so > quickly, I don''t have them any more. According to the FDSgraph charts > (which are messed up because I think it''s so busy, the graphs are just a > patchwork of points which aren''t joined up),Ok. Then, how about the error log?> it does about 70k ops per > minutes, peaking at about 140k.What are the clients? OS login?> This suddenly shot up a few weeks ago > and nobody seems to know why so I''m assuming that the load is the > problem. I was thinking about increasing the mem cache for FDS - worth > trying? It seems to be at the default (10 Meg?) and this box has easily > 3Gb doing nothing at any one time. >The mem cache should only be large enough to hold all of the entries/indexes in your database - any more than that is not used. You could try increasing it, but I doubt it will help.> PK > > -- > Fedora-directory-users mailing list > Fedora-directory-users@redhat.com > https://www.redhat.com/mailman/listinfo/fedora-directory-users >