On Mon, 2009-09-28 at 10:52 +0200, Claude Paroz wrote: > Hi, > > You may know that progress (l10n.gnome.org) is down from Friday. I'm a > little sad to see that such a capital application for translators can be > down for several days. > What's the status of this issue? What can I do? Please, give us > information... > > I don't want to accuse anyone, I know most of the community work as > volunteer, but I wonder if we can do something so as such a situation > don't arise again in the future. > > Should we move this application on another infrastructure, more > accessible to sysadmins? > Should we have more formal procedures for server upgrades?
James Troup, one of the Canonical sysadmins was around on #sysadmin and rebooted the box after turning off nss-ldap. I was able to log in. When I turned on nss-ldap again, things were OK for a while, but then slapd started running at 200% and no NSS lookups were succeeding making the box unresponsive in many ways. Stracing the slapd process showed it doing no IO but just sitting there churning - I don't know if we have some rogue client doing really expensive LDAP queries or if it's a slapd bug. I: - Turned off nss-ldap again and left it off - Added entries for Claude to /etc/passwd,/etc/shadow,/etc/group - Put an entry for gnomeweb in /etc/group with Claude as the only member. So that Claude could work on damned-lies. Someone else will have to investigate the slapd situation further. - Owen _______________________________________________ gnome-infrastructure mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnome-infrastructure
