Re: ASM Disgroup Mount
From: Sanjay Mishra <smishra_97_at_yahoo.com>
Date: Mon, 09 Feb 2015 21:29:52 +0000
Message-ID: <1548124338.2178189.1423517391443.JavaMail.yahoo_at_mail.yahoo.com>
Mladen
Recovery is not an issue but timing is big issue here. Currently already moved to Standby and taking backup to restore on primary. No San replication is used and so rman restore is will 2-3 days based on existing processing. Want to check if somehow it can be recovered with small recovery. Based on checking all looks goods. Disk reported can be acceible with dd and also can see the header is correct showing all detail with kfed. Question is if Oracle also maintains some Disk information used in ASM diskgroup outside of Disk header like in OCR and use it while mounting it. TxSanjay
Date: Mon, 09 Feb 2015 21:29:52 +0000
Message-ID: <1548124338.2178189.1423517391443.JavaMail.yahoo_at_mail.yahoo.com>
Mladen
Recovery is not an issue but timing is big issue here. Currently already moved to Standby and taking backup to restore on primary. No San replication is used and so rman restore is will 2-3 days based on existing processing. Want to check if somehow it can be recovered with small recovery. Based on checking all looks goods. Disk reported can be acceible with dd and also can see the header is correct showing all detail with kfed. Question is if Oracle also maintains some Disk information used in ASM diskgroup outside of Disk header like in OCR and use it while mounting it. TxSanjay
On Monday, February 9, 2015 2:58 PM, Mladen Gogala <dmarc-noreply_at_freelists.org> wrote:
On 02/09/2015 09:54 AM, Sanjay Mishra (Redacted sender smishra_97_at_yahoo.com for DMARC) wrote:
Getting issue on AIX 11.2.0.3 diskgroup
ORA-15032: not all alterations performed ORA-15040: diskgroup is incomplete ORA-15042: ASM disk "2" is missing from group number "3"ERROR: ALTER DISKGROUP DATADG MOUNT /* asm agent *//* Checked Disk_number 2 is corresponding to /dev/rhdisk23 and also v$asm_disk is showing it with Header_status MEMBER. Issue started when someone by mistake added wrong OCR disk into the DATADG and cause the crash. Wrong disk was removed at that time along with one more disk 2 but after date OCR was restored from backup, it look like Oracle might be reading Disk from some OCR metadata or somewhere and looking for Disk 2. As this is 30Tb Diskgroup and so want to bring the Diskgroup up with removing any disk and then recover the database instead of restoing 30T which will take days. Permission on the Disks are correct on all nodes along with Owner. Environment also has 6 more diskgroup and all of them are up and has corresponding database running. So need to know as what other steps that can be done to bring diskgroup up with even removing this affected diskgroup. Is this information about the disk are stored in any other place like OCR and as OCR is restore where disk 2 was removed earlier before the restored OCR back is issue TIA Sanjay
Hopefully, you have a standby database? Running 30 TB database without a standby would be irresponsible. If so, fail over to standby, that's why it's there, restore the database and recover it. This is where SAN snapshots can be very handy indeed.
--
Mladen Gogala
Oracle DBA
http://mgogala.freehostia.com
-- http://www.freelists.org/webpage/oracle-lReceived on Mon Feb 09 2015 - 22:29:52 CET