I'm trying some scenarios to see how syncrepl behaves in certain error conditions. In one of these slapd blocks and is terminated(?) (2.4.23 and 2.4.26, running on RH 5.5):

slapd D ffff810002536420 0 21032 1 21037 4521 (NOTLB) slapd D ffff810002536420 0 21032 1 21037 4521 (NOTLB) ffff810003f3fce8 0000000000000082 ffff81007eaae080 ffff81007faec100 000ec08328ac597e 0000000000000007 ffff81007eaae080 ffffffff80311b60 000ec0832bef5b8d 0000000000000990 ffff81007eaae268 0000000000000001 Call Trace: [<ffffffff8006315f>] wait_for_completion+0x79/0xa2 [<ffffffff8008e430>] default_wake_function+0x0/0xe [<ffffffff8004213e>] exit_mm+0x76/0xf5 [<ffffffff800158e6>] do_exit+0x305/0x965 [<ffffffff80049392>] cpuset_exit+0x0/0x88 [<ffffffff8002b42d>] get_signal_to_deliver+0x465/0x494 [<ffffffff8005af5e>] do_notify_resume+0x9c/0x7af [<ffffffff8008e430>] default_wake_function+0x0/0xe [<ffffffff80031545>] do_fork+0x148/0x1c1 [<ffffffff800a6061>] sys_futex+0x10a/0x12b [<ffffffff8005d19f>] sysret_signal+0x1c/0x27 [<ffffffff8005d427>] ptregscall_common+0x67/0xac INFO: task slapd:21037 blocked for more than 120 seconds.

How to reproduce:

Configuration: one master with a push replica (see also ITS 6999).

On the replica: stop slapd, clean out the database directory, set up the essential files, start slapd.

On the master: add a new entry to LDAP and run ldapsearch on that entry afterwards: the command hangs.

ldap log entries for this: conn=1002 fd=20 ACCEPT from PATH=/var/run/ldapi (PATH=/var/run/ldapi) conn=1002 op=0 BIND dn="cn=Manager" method=128 conn=1002 op=0 BIND dn="cn=Manager" mech=SIMPLE ssf=0 conn=1002 op=0 RESULT tag=97 err=0 text= conn=1002 op=1 ADD dn="uid=user2002,ou=People,dc=example,dc=com" slap_queue_csn: queing 0x43224150 20110831224446.739440Z#000000#001#000000 syncprov_sendresp: cookie=rid=001,sid=001,csn=20110831224446.739440Z#000000#001#000000 do_syncrep2: rid=001 cookie=rid=001,sid=001,csn=20110831224446.739440Z#000000#001#000000 syncrepl_entry: rid=001 LDAP_RES_SEARCH_ENTRY(LDAP_SYNC_ADD) conn=1002 op=2 UNBIND conn=1002 op=1 RESULT tag=105 err=0 text= slap_graduate_commit_csn: removing 0x2aaab0004160 20110831224446.739440Z#000000#001#000000 conn=1002 fd=20 closed syncrepl_entry: rid=001 be_search (0) syncrepl_entry: rid=001 uid=user2002,ou=People,dc=example,dc=com slap_queue_csn: queing 0x5d4ea70 20110831224446.739440Z#000000#001#000000 syncrepl_entry: rid=001 be_add uid=user2002,ou=People,dc=example,dc=com (32)

Is this an error in slapd or in RH?