On Sun, Dec 24, 2006 at 09:08:41AM -0500, Owen Taylor wrote: > On Sat, 2006-12-23 at 11:47 +0100, Olav Vitters wrote: > > On Fri, Dec 22, 2006 at 07:18:52PM +0000, Ross Golder wrote: > > > master, so I made label the master instead of button, and set up > > > replication from label to socket. I transferred the live configuration > > > and data from button to label, tested it, updated the 'ldap-back' > > > > Not having ldap on button improved the performance of Bugzilla > > significantly. So, thanks. :) > > That doesn't make sense to me ... performance should be about the > same whether LDAP is running remotely on label or remotely on button. > Label is a more powerful machine, but not significantly more so.
I actually was not thinking about the newer LDAP version. > And why would bugzilla be doing more than a very few lookups against > LDAP? (Especially if we are running nscd, which we probably should be.) The only LDAP stuff is looking up the home directory. Button used to be hovering around load 2+ (I thought higher, but the graphs do not show that). My guess is that because of that service being moved, button can allocate more time to MySQL. Strangely, the load on label even with openldap is much less than it used to be on button. Maybe because of the incomplete config mentioned by ninja. > We simply need to get the other machines upgraded ... migrating > everything to box and label because they are RHEL4 isn't sustainable. Agreed; I do not want anything else on box. Further, even box does not meet the minimum requirements for Bugzilla 3.0. -- Regards, Olav _______________________________________________ Gnome-infrastructure mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnome-infrastructure
