Mon Sep 29 11:35:22 PDT 2008
- Previous message: [Slony1-general] Time partitioned tables under slony
- Next message: [Slony1-general] futex_wait normal?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi, I'm noticing that there's a very old row in sl_confirm on one of my nodes. Wondering what could be the problem... test_slony_state says the following (I can paste more if needed): > Listen Path Analysis > =================================================== > No problems found with sl_listen > > -------------------------------------------------------------------------------- > Summary of event info > Origin Min SYNC Max SYNC Min SYNC Age Max SYNC Age > = > = > = > = > = > = > = > = > = > = > ====================================================================== > 1 5269981 5272295 00:00:00 00:20:00 0 > 4 1172599 1172992 00:00:00 00:20:00 0 > 2 846926 847325 00:00:00 00:20:00 0 > 3 1697518 1700013 00:00:00 00:21:00 0 > > > --------------------------------------------------------------------------------- > Summary of sl_confirm aging > Origin Receiver Min SYNC Max SYNC Age of latest SYNC Age > of eldest SYNC > = > = > = > = > = > = > = > = > = > = > = > ====================================================================== > 1 2 5269992 5272295 00:00:00 > 00:20:00 0 > 1 3 5269984 5272295 00:00:00 > 00:20:00 0 > 1 4 5269981 5272293 00:00:00 > 00:20:00 0 > 2 1 846926 847325 00:00:00 > 00:20:00 0 > 2 3 846926 847325 00:00:00 > 00:20:00 0 > 2 4 846926 847325 00:00:00 > 00:20:00 0 > 3 1 3778920 3778920 44 days 14:13:00 44 > days 14:13:00 1 > 3 2 1697665 1700013 00:00:00 > 00:20:00 0 > 3 4 1697518 1699955 00:00:00 > 00:21:00 0 > 4 1 1172599 1172992 00:00:00 > 00:20:00 0 > 4 2 1172599 1172992 00:00:00 > 00:20:00 0 > 4 3 1172599 1172992 00:00:00 > 00:20:00 0 > > > ------------------------------------------------------------------------------ > > Listing of old open connections > Database PID User Query > Age Query > = > = > = > = > = > = > = > = > = > = > ====================================================================== > > ISSUES FOUND: > > > Node: 3 Confirmations not propagating from 3 to 1 > ================================================ > Confirmations not propagating quickly in sl_confirm - > > For origin node 3, receiver node 1, earliest propagated > confirmation has age 44 days 14:13:00 > 00:30:00 > > Are slons running for both nodes? > > Could listen paths be missing so that confirmations are not > propagating? I've looked at sl_log1, sl_log2, and sl_event, but nothing corresponds to con_seqno in the sl_confirm table: > tii=# select * from _tii.sl_confirm order by con_timestamp; > con_origin | con_received | con_seqno | con_timestamp > ------------+--------------+-----------+---------------------------- > 3 | 1 | 3778920 | 2008-08-15 21:12:43.108286 > > tii=# select * from _tii.sl_event where ev_seqno = 3778920; > ev_origin | ev_seqno | ev_timestamp | ev_minxid | ev_maxxid | > ev_xip | ev_type | ev_data1 | ev_data2 | ev_data3 | ev_data4 | > ev_data5 | ev_data6 | ev_data7 | ev_data8 > -----------+----------+--------------+-----------+----------- > +--------+---------+----------+----------+----------+---------- > +----------+----------+----------+---------- > (0 rows) I have also restarted the slon daemons on nodes 1 and 3, but that doesn't change anything. Wondering if this may be the cause of erratic sync behavior? My monitoring indicates that this node (node 3) is regularly having lag issues, but the other two nodes (nodes 2 and 4) haven't even a squeak. Any ideas where else to look? What else can I do? --Richard
- Previous message: [Slony1-general] Time partitioned tables under slony
- Next message: [Slony1-general] futex_wait normal?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list