Hi,
last ween i've seen a blocking "automatic vacuum".
as i understood, this is not supposed to happen. in the past i saw vacuum processes disappear, in case of the need of a lock.
this is the relvant extract from the log and the database:
2010-06-18 18:33:36.011 CEST 172.19.5.34(57414) gc:29274 143897560 LOG: process 29274 still waiting for RowExclusiveLock on relation 42964239 of database 19759903 after 1002.601 ms
...
2010-06-18 19:23:43.898 CEST :20892 0 LOG: automatic vacuum of table "gd.pg_toast.pg_toast_42964236": index scans: 1
pages: 1469113 removed, 288899 remain
tuples: 369645 removed, 396719 remain
system usage: CPU 5.46s/0.85u sec elapsed 13785.76 sec
2010-06-18 19:23:43.899 CEST 172.19.5.22(45561) gc:315 143897590 LOG: process 315 acquired RowExclusiveLock on relation 42964239 of database 19759903 after 3007903.169 ms
2010-06-18 19:23:43.920 CEST 172.19.5.22(45561) gc:315 143897590 LOG: duration: 3007924.883 ms execute <unnamed>: insert into SHARED_GAMESET .....
gd=# select oid,* from pg_class where reltoastrelid=42964239;
relname | relnamespace | reltype | relowner | relam | relfilenode | reltablespace | relpages | reltuples | reltoastrelid | reltoastidxid | relhasindex | relisshared | relistemp | relkind | relnatts | relchecks | rel
hasoids | relhaspkey | relhasrules | relhastriggers | relhassubclass | relfrozenxid | relacl | reloptions
----------------+--------------+----------+----------+-------+-------------+---------------+----------+-------------+---------------+---------------+-------------+-------------+-----------+---------+----------+-----------+----
--------+------------+-------------+----------------+----------------+--------------+----------------------------------------+------------
shared_gameset | 19760303 | 42964238 | 16443 | 0 | 42964236 | 0 | 143715 | 1.96642e+06 | 42964239 | 0 | t | f | f | r | 4 | 0 | f
| t | f | f | f | 101290258 | {gdadm=arwdDxt/gdadm,gc=arwdDxt/gdadm} |
(1 row)
So what's this?
from my point of view the vacuum blocked inserts on this table for about 50 minutes.
Is this a bug?
Or do yuo see here a configuration issue?
best regards,
Uwe
Uwe Bartels
Systemarchitect - Freelancer
mailto: uwe.bartels at gmail.com
