Error 10612 on Replication Source DB

Posted by James Palmer on 17-Jul-2017 16:31

Progress 10.2B08 on Windows. Can't start up a DB with following in log file:

[2017/07/17@22:28:59.486+0100] P-7764 T-7704 F : (10612) SYSTEM ERROR: numlive 1 tx_count 1
logTxEntry transnum 945924525 rlcounter 66078
pTranEntry transnum 0 rlcounter 0
rlmemchk: transaction table mismatch.

Any ideas what we can do to get the DB started? Even if it's just a case of getting access to get the data out to repopulate a new DB? We've already put an empty DB in place so they can carry on operations. 

All Replies

Posted by Mike Fechner on 17-Jul-2017 16:37

You have tried -F already?

Posted by James Palmer on 17-Jul-2017 16:38

Thanks Mike - leaving that for a last resort.

Posted by Brian Bowman on 17-Jul-2017 16:44

Is this the same machine that you had the 2 replication target problems with out of the 10 databases on it?
 
If yes then something sounds not quite right on the machine…
 
Brian
 
Brian L. Bowman
 
Senior Principal Product Manager
Progress Software Corporation
14 Oak Park, Bedford, MA, USA 01730
 
Phone: +1 (603) 801-8259
Email: bowman@progress.com
 
500x100 Email Signature
 

Posted by James Palmer on 17-Jul-2017 16:52

It's the replication source for the replication target we had problems with earlier Brian. Something is definitely not right. They've been doing something funky with snapshots and virtualisation from what I can tell. We've recovered the main database with just index corruption, but this is the audit database and it's totally hosed.

Posted by George Potemkin on 17-Jul-2017 22:55

> logTxEntry transnum 945924525 rlcounter 66078

945924525 = 0x3861A9AD

These four bytes look like a text fragment - "8a" + two non-asci characters.

Maybe the wrong data was written to the bi file. For example, a wrong file handle was used to write data block. If it's true then bi file is lost and the -F option is the only choice.

What happened with database before it's crashed? Memory violation errors?

This thread is closed