Hi,
we've migrated our production systems to 389ds version 1.3.2.24. Everything seems fine now, the only new messages i see in error logs (several times per day) are
[15/Nov/2014:03:58:43 +0100] - replica_generate_next_csn: opcsn=5466c164000000010000 <= basecsn=5466c164000000020000, adjusted opcsn=5466c164000100010000 [15/Nov/2014:10:38:38 +0100] - replica_generate_next_csn: opcsn=54671f1f000000010000 <= basecsn=54671f1f000000030000, adjusted opcsn=54671f1f000100010000
Are these only information messages that can be safely ignored or they may be a manifestation of some potential problem? In source code (./ldap/servers/plugins/replication/repl5_replica.c) it looks like a serious one (SLAPI_LOG_FATAL): slapi_log_error (SLAPI_LOG_FATAL, NULL, "replica_generate_next_csn: " "opcsn=%s <= basecsn=%s, adjusted opcsn=%s\n", opcsnstr, basecsnstr, opcsn2str);
Thanks!
On 11/15/2014 10:30 AM, Ivanov Andrey (M.) wrote:
Hi,
we've migrated our production systems to 389ds version 1.3.2.24. Everything seems fine now, the only new messages i see in error logs (several times per day) are
[15/Nov/2014:03:58:43 +0100] - replica_generate_next_csn: opcsn=5466c164000000010000 <= basecsn=5466c164000000020000, adjusted opcsn=5466c164000100010000 [15/Nov/2014:10:38:38 +0100] - replica_generate_next_csn: opcsn=54671f1f000000010000 <= basecsn=54671f1f000000030000, adjusted opcsn=54671f1f000100010000
Are these only information messages that can be safely ignored or they may be a manifestation of some potential problem?
This looks ok to me, and the message should not be a fatal message. The code handles this correctly by incrementing the sequence number and updating the generator. In practice it should be very difficult to get the generator to generate a CSN like this. Are all of these machines running in VMs? If so, what is the hypervisor? How many of these do you see per day?
In source code (./ldap/servers/plugins/replication/repl5_replica.c) it looks like a serious one (SLAPI_LOG_FATAL): slapi_log_error (SLAPI_LOG_FATAL, NULL, "replica_generate_next_csn: " "opcsn=%s <= basecsn=%s, adjusted opcsn=%s\n", opcsnstr, basecsnstr, opcsn2str);
It should not be FATAL. Please file a ticket.
Thanks!
-- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
[15/Nov/2014:03:58:43 +0100] - replica_generate_next_csn: opcsn=5466c164000000010000 <= basecsn=5466c164000000020000, adjusted opcsn=5466c164000100010000 [15/Nov/2014:10:38:38 +0100] - replica_generate_next_csn: opcsn=54671f1f000000010000 <= basecsn=54671f1f000000030000, adjusted opcsn=54671f1f000100010000
Are these only information messages that can be safely ignored or they may be a manifestation of some potential problem?
This looks ok to me, and the message should not be a fatal message. The code handles this correctly by incrementing the sequence number and updating the generator.
That's what i've also thought.
In practice it should be very difficult to get the generator to generate a CSN like this. Are all of these machines running in VMs? If so, what is the hypervisor? How many of these do you see per day?
We see it two or three time per day, compared to 20000 or 30000 modifications per day (according to logconv.pl) : 38465 2.16.840.1.113730.3.5.12 DS90 Start Replication Request 24603 2.16.840.1.113730.3.5.5 End Replication Request (incremental update)
2 servers are physical (replica id 1 and 3) and one is virtual (replica id 2). Each of the three is MMR-replicated to two others. On rep_id 1 (physical hardware): [15/Nov/2014:03:58:43 +0100] - replica_generate_next_csn: opcsn=5466c164000000010000 <= basecsn=5466c164000000020000, adjusted opcsn=5466c164000100010000 [15/Nov/2014:10:38:38 +0100] - replica_generate_next_csn: opcsn=54671f1f000000010000 <= basecsn=54671f1f000000030000, adjusted opcsn=54671f1f000100010000 [16/Nov/2014:01:43:44 +0100] - replica_generate_next_csn: opcsn=5467f341000000010000 <= basecsn=5467f341000100020000, adjusted opcsn=5467f341000100010000 [17/Nov/2014:09:34:54 +0100] - replica_generate_next_csn: opcsn=5469b32f000000010000 <= basecsn=5469b32f000000020000, adjusted opcsn=5469b32f000100010000 [17/Nov/2014:16:09:48 +0100] - replica_generate_next_csn: opcsn=546a0fbd000000010000 <= basecsn=546a0fbd000200020000, adjusted opcsn=546a0fbd000300010000 [17/Nov/2014:16:55:55 +0100] - replica_generate_next_csn: opcsn=546a1a8c000000010000 <= basecsn=546a1a8c000000020000, adjusted opcsn=546a1a8c000100010000 [17/Nov/2014:19:34:14 +0100] - replica_generate_next_csn: opcsn=546a3fa7000000010000 <= basecsn=546a3fa7000000030000, adjusted opcsn=546a3fa7000100010000
On rep_id 2 (virtual, VMWare ESXi5.5): [15/Nov/2014:04:19:09 +0100] - replica_generate_next_csn: opcsn=5466c62e000000020000 <= basecsn=5466c62e000000030000, adjusted opcsn=5466c62e000100020000 [17/Nov/2014:15:47:11 +0100] - replica_generate_next_csn: opcsn=546a0a70000100020000 <= basecsn=546a0a70000200030000, adjusted opcsn=546a0a70000200020000 [17/Nov/2014:15:48:11 +0100] - replica_generate_next_csn: opcsn=546a0aac000100020000 <= basecsn=546a0aac000200030000, adjusted opcsn=546a0aac000200020000 [17/Nov/2014:15:49:36 +0100] - replica_generate_next_csn: opcsn=546a0b01000000020000 <= basecsn=546a0b01000200030000, adjusted opcsn=546a0b01000300020000
On rep_id 3 (physical hardware): [16/Nov/2014:05:02:34 +0100] - replica_generate_next_csn: opcsn=546821db000000030000 <= basecsn=546821dc000000020000, adjusted opcsn=546821dc000100030000
In source code (./ldap/servers/plugins/replication/repl5_replica.c) it looks like a serious one (SLAPI_LOG_FATAL): slapi_log_error (SLAPI_LOG_FATAL, NULL, "replica_generate_next_csn: " "opcsn=%s <= basecsn=%s, adjusted opcsn=%s\n", opcsnstr, basecsnstr, opcsn2str);
It should not be FATAL. Please file a ticket.
Ok. Done: https://fedorahosted.org/389/ticket/47959
Thanks!
389-users@lists.fedoraproject.org
-
Ivanov Andrey (M.) -
Rich Megginson