21 Mar
2009
21 Mar
'09
2:22 p.m.
Aaron Richton wrote:
Please see https://www.nbcs.rutgers.edu/~richton/testrun.20090321.tgz (includes two backtraces too). I'm not sure if the slapd logs are enough to determine sufficient mutex details; if they're not, I think I can play around with dbx a bit.
I presume that the lockup is in server 3 / slapd.3.log ... ?
In backtrace.0 there are two threads t@6 and t@7 spinning on ldap_pvt_thread_yield, can you go up one frame and print si, *si in each thread?
--
-- Howard Chu
CTO, Symas Corp. http://www.symas.com
Director, Highland Sun http://highlandsun.com/hyc/
Chief Architect, OpenLDAP http://www.openldap.org/project/