Corrupted Log File

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Corrupted Log File

greenbean
I have noticed this error a few times.  It occurs very infrequently.  Will this cause any problems with a running system, or just prevent proper replays if the system happens to go down when a log file is corrupted?
 
bitronix.tm.journal.CorruptedTransactionLogException: corrupted log found at position 1500554 (record terminator outside of file bounds: 1500623 of 1500603, recordLength: 61)
        at bitronix.tm.journal.TransactionLogCursor.readLog(TransactionLogCursor.java:91)
        at bitronix.tm.journal.TransactionLogCursor.readLog(TransactionLogCursor.java:65)
        at bitronix.tm.journal.DiskJournal.collectDanglingRecords(DiskJournal.java:359)
        at bitronix.tm.journal.DiskJournal.collectDanglingRecords(DiskJournal.java:204)
        at bitronix.tm.recovery.Recoverer.run(Recoverer.java:135)
        at java.lang.Thread.run(Thread.java:662)
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Corrupted Log File

Ludovic Orban-2
This exception shows that the background recoverer is working on corrupted files. That's reasonably okay as the background recoverer will alway run at a later time and the journal never discards valuable recovery information.

The first idea anyone would have is to think that there is a bug in the disk journal code. While this is possible, it's _extremely_ unlikely as the disk journal has been left nearly untouched for most of BTM's life and I've never observed nor heard about corrupted journal files which wouldn't panic the TM after a short period of time.

What I suspect is that you have a hardware problem with your disks. Unless you're storing the BTM journal files on a fiber-channel SAN, or on an ultra-safe file system like ZFS I would try out a different hardware setup.

If you're convinced your disks aren't at fault, then we'll need to have a closer look at the problem.


2011/3/3 Urciolo, Kevin J (IS) <[hidden email]>
I have noticed this error a few times.  It occurs very infrequently.  Will this cause any problems with a running system, or just prevent proper replays if the system happens to go down when a log file is corrupted?
 
bitronix.tm.journal.CorruptedTransactionLogException: corrupted log found at position 1500554 (record terminator outside of file bounds: 1500623 of 1500603, recordLength: 61)
        at bitronix.tm.journal.TransactionLogCursor.readLog(TransactionLogCursor.java:91)
        at bitronix.tm.journal.TransactionLogCursor.readLog(TransactionLogCursor.java:65)
        at bitronix.tm.journal.DiskJournal.collectDanglingRecords(DiskJournal.java:359)
        at bitronix.tm.journal.DiskJournal.collectDanglingRecords(DiskJournal.java:204)
        at bitronix.tm.recovery.Recoverer.run(Recoverer.java:135)
        at java.lang.Thread.run(Thread.java:662)