[Date Prev][Date Next] [Chronological] [Thread] [Top]

RE: HEADS UP: slapd outbound connections



> -----Original Message-----
> From: Quanah Gibson-Mount [mailto:quanah@stanford.edu]

> > Unfortunately, this appears to cause the replica's using syncRepl to
> > segfault:
>
> I'll confirm syncRepl replica's only.  My non-syncRepl host with same
> checkout works fine on queries.

I was able to reproduce the crash, it was due to the connection idletimeout
handler. I changed that to ignore outbound connections. This revealed a
couple other loose ends in the error recovery handling and runqueue
scheduling. I think it's stable now: if a persistent session dies
(LDAP_SERVER_DOWN), it will reschedule itself to retry, instead of just
quitting.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support