MySQL logfile questions

[Marty]

I've been looking at my MYSQL_ROOT:[MYSQL_SERVER]MYSQLD.LOG files and have noticed that most of them report that the DB didn't shutdown correctly and are chock full of read error messages...

Code: Select all

...logfile snippet...
$ mysqld  --default-storage-engine=InnoDB --innodb-safe-binlog --ansi --myisam-
060524  7:40:07  InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: cannot read directory ./joomla, error 17
InnoDB: Error: os_file_readdir_next_file() returned -1 in
InnoDB: directory ./joomla

...this goes on for a while, through different directories, then...

Code: Select all

InnoDB: Crash recovery may have failed for some .ibd files!
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
060524  7:40:11  InnoDB: Starting log scan based on checkpoint at
InnoDB: log sequence number 0 314459.
InnoDB: Doing recovery: scanned up to log sequence number 0 314459
InnoDB: Last MySQL binlog file position 0 1494, file name ./black-bin.000022
060524  7:40:12  InnoDB: Flushing modified pages from the buffer pool...
060524  7:40:12  InnoDB: Started; log sequence number 0 314459
060524  7:40:12 [ERROR] After InnoDB crash recovery, checking if the
binary log './black-bin.000022' contains rolled back transactions
which must be removed from it...
/$38$DKA300/MYSQL/VMS/BIN/mysqld.exe: ready for connections.
Version: '4.1.14-log'  socket: ''  port: 3306  Source distribution

What does this all mean? If I look at the tail of the previous logfile it seems to have shutdown correctly...

Code: Select all

060523 22:39:18 [Note] /$38$DKA300/MYSQL/VMS/BIN/mysqld.exe: Normal shutdown
060523 22:40:48  InnoDB: Starting shutdown...

[WillemGrooters]

060523 22:39:18 [Note] /$38$DKA300/MYSQL/VMS/BIN/mysqld.exe: Normal shutdown
060523 22:40:48 InnoDB: Starting shutdown...

It would have been the right thing to do to signal shutwdown had finished, I think. So if InnoDB does NOT signal that situation, it can mean three things:
* it DID finish properly but had the bad habit not to signal it
* it did NOT finish properly, and had the bad habit not to signal it
* it crashed in shutting down

Given the responses at startup, I think the second or third has been the case. Perhaps you can find more in Accounting.