[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Re: (sigh) replication still failing on 1.2.13
After augmenting the fopen_lock function in servers/slapd/lock.c with a
call to strerror(), the syslog messages changes to:
Feb 28 12:33:59 ns01 slapd[364]: conn=24314 op=1 MOD
dn="UID=ASPARKS,OU=PEOPLE,DC=QURIS,DC=COM"
Feb 28 12:34:00 ns01 slapd[364]: conn=24314 op=1 RESULT err=0 tag=103
nentries=0Feb 28 12:34:00 ns01 slapd[364]: could not open
"/opt/ldap/var/replogs/dc=quris,dc=com.lock":Error 0
Error 0... Not anywhere nearly helpful enough...
I note the Solaris man page for fopen includes the following:
"The fopen() function may fail and not set errno if there are
no free stdio streams."
Don't understand why I wouldn't have enough, or how to increase the # of
streams. File descriptors, I do, and have set that (in /etc/system and
in a ulimit before starting slapd) to 1024.
Anyone have any advice? Thanks in advance.
-Alan
Alan Sparks wrote:
>
> Well, my problem with replication on OL 1.2.13 with Solaris7 isn't
> threads, after all. The non-threaded version is also failing...
>
> I bumped up debug logging, and wrote a couple of Perl progs to force
> something to change and check the replication. Finally failed today
> with the following logged:
>
> Feb 27 17:36:46 ns01 slapd[20060]: conn=50508 op=1 MOD
> dn="UID=ASPARKS,OU=PEOPLE,DC=QURIS,DC=COM"
> Feb 27 17:36:46 ns01 slapd[20060]: conn=50508 op=1 RESULT err=0 tag=103
> nentries=0
> Feb 27 17:36:46 ns01 slapd[20060]: could not open
> "/opt/ldap/var/replogs/dc=quris,dc=com.lock"
>
> Unfortunately, slapd doesn't log the errno, so who knows why. I'm
> modifying the code to include a strerror() in the message to find out
> why.
> I set the ulimit up to 1024 file descriptors before starting slapd. I
> sure hope I'm not hitting that, can't see why unless something is
> leaking fd's.
>
> -Alan
>
> --
> Alan Sparks, Sr. UNIX Administrator asparks@quris.com
> Quris, Inc. (720) 836-2058
--
Alan Sparks, Sr. UNIX Administrator asparks@quris.com
Quris, Inc. (720) 836-2058