Informix - Problem description
Problem IT27884 | Status: Closed |
ON AN UPDATABLE SECONDARY, THE ROLLBACK OF A LARGER TRANSACTION CAN BE VERY SLOW | |
product: | |
INFORMIX SERVER / 5725A3900 / C10 - IDS 12.10 | |
Problem description: | |
If a larger transaction - allocating many locks - is replicated to an updatable secondary (updating as many locks there too), but eventually the transaction gets rolled back, depending on size of the lock list this rollback can be extremely slow on the secondary, potentially also involving DR disconnects and persistent reconnect failure. Symptoms are: - many locks allocated on secondary - current log position moving very slowly in: onstat -l -r 1 | grep C- - an xchg_1. thread running pretty much permanently - onlog showing large amounts of CLR log records at current log position | |
Problem Summary: | |
**************************************************************** * USERS AFFECTED: * * Users of IDS prior to 12.10.xC13. * **************************************************************** * PROBLEM DESCRIPTION: * * On an updatable secondary, the rollback of a larger * * transaction can be very slow. * **************************************************************** * RECOMMENDATION: * **************************************************************** | |
Local Fix: | |
What can help overcome such situation is restarting the secondary as non-updatable. To avoid it up front: obtain an exclusive table lock before running large jobs | |
Solution | |
Workaround | |
not known / see Local fix | |
Timestamps | |
Date - problem reported : Date - problem closed : Date - last modified : | 25.01.2019 24.09.2019 24.09.2019 |
Problem solved at the following versions (IBM BugInfos) | |
12.10.xC13 | |
Problem solved according to the fixlist(s) of the following version(s) |