We have a database that has been set up with 2 ASM diskgroups: DATA and ONLINELOG. The ONLINELOG diskgroup holds all the archive redo logs and some online redo log members. During the weekend gone the alert log complained that the ONLINELOG diskgroup disk space had been exhausted whilst Oracle was trying to archive logs. This eventually led to the ASM instance being terminated.
Wehave managed to get to database back up in mount state. However we cannot mount the ONLINELOG diskgroup as Oracle is complaining of a disk write lost failure (ORA-15096: lost disk write detected ). This means that a failure either by disk hardware or disk software caused a disk write to be lost, even though ASM received acknowledgement that the write completed.
The official solution mentions that the diskgroup is corrupt and cannot be recovered. The disk group must be recreated, and its contents restored from backups. We donít have any backups (this solution is not managed by us, we are trying to help another part of the company). Hence, my question here is there anyway we can open the database? (There are some members of the online redo logs on the available diskgroup)
I think 2/3 groups are available on the available diskgroup but the 'current' logfile is in the damaged diskgroup. Is it possible to recreate the control file and just point to the logs available and try to open the database?
Ahh the most important thing is before you attempt any recovery
a)Keep a CoolHead dont panic
b)take a backup of whatever is remaining so in case you screw up you get back to whateever state you were before the screw up :-)
c)Contact oracle support before attempting a recovery after all thats why we pay oracle support for ;-)
Last resort you need to try something called as oracle DuDe which can extract data from a database like this
Yes they are taking a backup
They have contacted Oracle support and they have mentioned force opening the database using a hidden parameter which may allow for corruption. Once open, export out the data needed. This is their case might not be feasiable time wise due to the huge amounts of data in the tables
If they manage to open the database (using the hidden parameter), the application data is organised in weekly tablespaces with daily partitions. Hence if they do manage to open the db will they be able to take exports of weekly tablespaces from a certain date (as long as datafiles are online and consistent) and plug them into a new database? Under normal circumstances this would would work but as all the data is partitioned would the base tables have to exported as well?
Disclaimer: Advice is provided to the best of my knowledge but no implicit or explicit warranties are provided. Since the advisor explicitly encourages testing any and all suggestions on a test non-production environment advisor should not held liable or responsible for any actions taken based on the given advice.