Full_Name: Andreas Hasenack Version: 2.3.11 OS: linux URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (200.206.15.118) I have a 2.3.11 that is crashing during the night. To try to reproduce it, I ran some thousand queries (with START_TLS [-ZZ]) while the server was running with -d -1. After about 16k queries and a 950Mb log file, I got this assertion error: slapd: daemon.c:524: slapd_set_read: Assertion `((slap_daemon.sd_index[s]) != -1)' failed A little bit before this I have some TLS errors: tls_read: want=5 error=Resource temporarily unavailable (could be relevant or not). The server hardware is a Xeon 3.2GHz (HT) with 1.5Gb of RAM and U320 SCSI disks in mirror raid (IBM xseries 336). I'm running an SMP kernel 2.6.12-12mdksmp (stock from Mandriva 2006).
Got this from gdb: #0 0xffffe410 in __kernel_vsyscall () #1 0xb7aa3ef1 in raise () from /lib/tls/libc.so.6 #2 0xb7aa583b in abort () from /lib/tls/libc.so.6 #3 0xb7a9d045 in __assert_fail () from /lib/tls/libc.so.6 #4 0x08066473 in slapd_set_read (s=Variable "s" is not available.) at daemon.c:524 #5 0x0806d16b in connection_read (s=20) at connection.c:1345 #6 0x080670c7 in slapd_daemon_task (ptr=0x0) at daemon.c:1979 #7 0xb7d62c40 in start_thread () from /lib/tls/libpthread.so.0 #8 0xb7b480ee in clone () from /lib/tls/libc.so.6
On Thursday 03 November 2005 17:39, andreas@conectiva.com.br wrote: > Full_Name: Andreas Hasenack > Version: 2.3.11 > OS: linux > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (200.206.15.118) > > > I have a 2.3.11 that is crashing during the night. To try to reproduce it, > I ran some thousand queries (with START_TLS [-ZZ]) while the server was > running with -d -1. > After about 16k queries and a 950Mb log file, I got this assertion error: > > slapd: daemon.c:524: slapd_set_read: Assertion `((slap_daemon.sd_index[s]) > != -1)' failed I just ran into this, it is ITS 4108. I've attached my backport of the fix for 2.3.11, which has been running under substantial load for almost an hour now without triggering it. Without the patch I could trigger it under less load in under 3 minutes. > > A little bit before this I have some TLS errors: > tls_read: want=5 error=Resource temporarily unavailable > > (could be relevant or not). > > The server hardware is a Xeon 3.2GHz (HT) with 1.5Gb of RAM and U320 SCSI > disks in mirror raid (IBM xseries 336). I'm running an SMP kernel > 2.6.12-12mdksmp (stock from Mandriva 2006). Please try 2.3.11-6mdk once it is uploaded ... Regards, Buchan -- Buchan Milne ISP Systems Specialist B.Eng,RHCE(803004789010797),LPIC-2(LPI000074592)
On Thu, Nov 03, 2005 at 06:12:06PM +0200, Buchan Milne wrote: > I just ran into this, it is ITS 4108. I've attached my backport of the fix for > 2.3.11, which has been running under substantial load for almost an hour now > without triggering it. Without the patch I could trigger it under less load > in under 3 minutes. Applied, thanks. Running good so far. Previous version would stop after a few minutes of three parallel search scripts. > Please try 2.3.11-6mdk once it is uploaded ... I need my patched smbk5pwd, we can talk about this later with more time when I come back.
This problem should be fixed in OPENLDAP_REL_ENG_2_3. Please test.
changed state Open to Feedback moved from Incoming to Software Bugs
On Thu, Nov 03, 2005 at 06:55:15PM +0000, Kurt Zeilenga wrote: > This problem should be fixed in OPENLDAP_REL_ENG_2_3. Please test. Just retested here on an SMP machine, seems fine now.
changed notes
changed state Feedback to Release
changed notes changed state Release to Closed
moved from Software Bugs to Archive.Software Bugs
fixed in HEAD/RE23 dup of ITS#4108