AIX

Around a month ago we suffered a prolonged power outage, due to circumstances the servers and storage arrays were still in use when the UPS dropped. One of the servers was running a flash copy on the database while batch processing had commenced. This is the only server that suffered any further problems.

On restoration of the power, all systems returned. The databases recovered and were proved to be consistent, prior to system release to the users a backup was taken. At this point the system was released to the users and there was full normal operation of the database during the remainder of the online day - some 10 hours.

On starting batch processing that evening at 20:00 it was noticed that a number of the flash volumes were still in established relationships with the live volumes. These relationships could not be broken from the server and had to be broken at a SAN level, once broken the live volumes that were established became corrupt and also the database also became corrupt - requireing a restore from the backup.

This happened just over a month ago now, the system and database logs have been examined in detail without result. Although the systems have functioned without problem since the incident, there is still uncertainty about the cause - I have looked for information in manuals, the IBM web site and in several forums but without success

Has anyone seen anything like this before?

what type of SAN STORAGE you have ?

Sorry should have said,

The server is an LPAR on a p595 and attached storage is DS8300 fiber attached.

Regards