Hi guys,
I'm facing an issue regarding normal redundancy disk groups. Maybe I just did not understand correctly how ASM works - or it is just a minor configuration issue.
My system:
One database host (12.1, single instance) storage is on SAN. The storage shall be on ASM, to keep it simple I have has one LUN presented to my DB host. First I installed ASM, database and stuff with external redundancy (the SAN storage is mirroring the data). This setup is working as expected without trouble (i.e. create diskgroup data externalredundancy disk '/dev/sdc';).
Now my question: To increase redundancy my idea was to put another SAN storage (different datacenter reachable via different HBA from the same host with identical disk config) to have a simple mirror copy of the database in a 2nd location.
I already learned that one can not modify the redundancy level (here: from EXTERNAL to NORMAL) so I was going to drop & recreate the diskgroup.
Now the part I do not understand:
After backup up the database and re-creating the diskgroup (create diskgroup data normal redundancy failgroup FG1 disk '/dev/sdc' failgroup FG2 disk '/dev/sdd';) I was able to restore the database and the 2nd storage was also busy (blinking HDD LEDs) but I expect if the 2nd storage is failing completely (I disconnected the fibre cable between the data centers) the database will still be available - but it was not. The database went down and did not come up as long as I have not connected the 2nd storage again.
I was expecting the normal redundancy is protecting against exactly such a failure? And I'm pretty sure I did such an setup years ago with 11.1 (maybe 11.2) without trouble?
Also, I would expect "alter diskgroup data drop disks in failgroup FG2" would work - but is telling me it is not compatible with the redundancy level? Why in the world do I have two mirror copies?
Please can somebody tell me where my error is?
Message was edited by: user8946834
Fixed some typos