Thread 1 cannot allocate new log, sequence...

I sometimes expecting this message in alert.log file:
Thread 1 cannot allocate new log, sequence nnnnn

It can be skipped because it is not harmful but I decided to explore.

SQL> alter system set log_checkpoints_to_alert=true;

SQL> alter system checkpoint;

let's get some outcome at first.


Sometimes

Thread 1 cannot allocate new log, sequence

followed by

Private strand flush not complete

It seems to be quite enough to have 100Mb for one log:

 SQL> select group#, bytes, status from v$log;

    GROUP#      BYTES STATUS
---------- ---------- ----------------
         1  104857600 INACTIVE
         2  104857600 INACTIVE
         3  104857600 INACTIVE
         4  104857600 INACTIVE
         5  104857600 CURRENT

But :

 Redo size:     5,465.15 Per Second    25,738.80 Per Transaction

looks not  reassuring.

 Incremental checkpoint up to RBA [0x5b9d.e99b.0], current log tail at RBA [0x5b9d.1258c.0]
Tue Nov 23 09:36:51 2010
Incremental checkpoint up to RBA [0x5b9d.18d05.0], current log tail at RBA [0x5b9d.1a7df.0]
Tue Nov 23 09:41:52 2010
Thread 1 cannot allocate new log, sequence 23454
Private strand flush not complete
  Current log# 2 seq# 23453 mem# 0: /u1/oracle/XXX/ars/online/redo2a.ora
  Current log# 2 seq# 23453 mem# 1: /u2/oracle/XXX/ars/online/redo2b.log
LNS1 started with pid=35, OS id=11439
Tue Nov 23 09:41:57 2010
Beginning log switch checkpoint up to RBA [0x5b9e.2.10], SCN: 1178579247705
Thread 1 advanced to log sequence 23454
  Current log# 3 seq# 23454 mem# 0: /u1/oracle/XXX/ars/online/redo3a.ora
  Current log# 3 seq# 23454 mem# 1: /u2/oracle/XXX/ars/online/redo3b.log
Tue Nov 23 09:41:59 2010
LNS: Standby redo logfile selected for thread 1 sequence 23454 for destination LOG_ARCHIVE_DEST_2
Tue Nov 23 09:42:19 2010
ARC0: Standby redo logfile selected for thread 1 sequence 23453 for destination LOG_ARCHIVE_DEST_2
Tue Nov 23 09:46:56 2010
Completed checkpoint up to RBA [0x5b9e.2.10], SCN: 1178579247705
Tue Nov 23 10:06:59 2010
Incremental checkpoint up to RBA [0x5b9e.5233.0], current log tail at RBA [0x5b9e.7948.0]
Tue Nov 23 10:37:05 2010
Incremental checkpoint up to RBA [0x5b9e.13bae.0], current log tail at RBA [0x5b9e.160a7.0]
Tue Nov 23 11:07:12 2010
Incremental checkpoint up to RBA [0x5b9e.1eab3.0], current log tail at RBA [0x5b9e.20462.0]
Tue Nov 23 11:28:01 2010

looks like one long transaction step over itself.
I think it is only way is to make redo larger.