Re: Trying to Simulate a disk failure for one of the disks used by ASM disk group
- From: Hanan Hit <hithanan@xxxxxxxxx>
- To: Seth Miller <sethmiller.sm@xxxxxxxxx>
- Date: Thu, 28 Aug 2014 17:40:30 -0700
Hi All,
Thanks for the group assistance.
Seth had it
Here is what I have done
# echo offline > /sys/block/sdd/device/state
At this point I saw that the failure:
ASMCMD commands to gather complementary metadata information:
==================================
State Type Rebal Sector Block AU Total_MB Free_MB
Req_mir_free_MB Usable_file_MB Offline_disks Voting_files Name
MOUNTED NORMAL N 512 4096 4194304 8576580 1310328
571772 369278 1 N DATADG/
MOUNTED NORMAL N 512 4096 4194304 4574176 4567552
571772 1997890 0 N FRADG/
==================================
In order to recover from that on the Linux side just run
echo running > /sys/block/sdd/device/state
For “Fixing” the failure in the ASM instance just:
alter diskgroup datadg online disk DATA03;
Once the above is done DATADG no disk would be reported as failed.
Thx,
Hanan
On Aug 28, 2014, at 2:59 PM, Hanan Hit <hithanan@xxxxxxxxx> wrote:
> Ha Got it.
>
> Thanks for your help.
>
> Will try that and will keep you posted.
>
> Best,
> Hanan
>
> On Aug 28, 2014, at 2:57 PM, Seth Miller <sethmiller.sm@xxxxxxxxx> wrote:
>
>> The command I used would delete the "/dev/sdc" block device (along with its
>> partition devices) but you would replace that with whichever block device
>> you want to make go away.
>>
>> Seth Miller
>>
>>
>> On Thu, Aug 28, 2014 at 4:54 PM, Hanan Hit <hithanan@xxxxxxxxx> wrote:
>> I can I choose only a single drive - or maybe I didn’t understand that.
>>
>> Thanks,
>> Hanan
>>
>> On Aug 28, 2014, at 2:52 PM, Seth Miller <sethmiller.sm@xxxxxxxxx> wrote:
>>
>>> Hanan,
>>>
>>> You can simply delete the virtual representation of the device from the
>>> scsi subsystem.
>>>
>>> echo 1 > /sys/block/sdc/device/delete
>>>
>>> Seth Miller
>>>
>>>
>>> On Thu, Aug 28, 2014 at 1:34 PM, Mark W. Farnham <mwf@xxxxxxxx> wrote:
>>> doh. right you are.
>>>
>>>
>>>
>>> From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
>>> On Behalf Of Matthew Zito
>>> Sent: Thursday, August 28, 2014 1:08 PM
>>> To: Mark W. Farnham
>>> Cc: hithanan@xxxxxxxxx; Chitale, Hemant K; ORACLE-L
>>>
>>>
>>> Subject: Re: Trying to Simulate a disk failure for one of the disks used by
>>> ASM disk group
>>>
>>>
>>>
>>> Remember, it's ASM, so there's no mounting or unmounting!
>>>
>>>
>>> Changing the permissions *might* work, but on Linux, since you still do an
>>> open() and get a file descriptor even when you're doing direct I/O, I think
>>> it would bypass it if the database is already running (since it already has
>>> a valid FD it's writing to/from).
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Thu, Aug 28, 2014 at 1:03 PM, Mark W. Farnham <mwf@xxxxxxxx> wrote:
>>>
>>> If this is Linux or Unix, then probably umount followed by a mount readonly
>>> would do the trick if you’re writing to that disk at all.
>>>
>>>
>>>
>>> Possibly changing the permissions would intervene, but I think that varies
>>> about whether that will stop a running application that already has a file
>>> open.
>>>
>>>
>>>
>>> Heh. It was easier when there was a button on each drive you could toggle
>>> to make it read only.
>>>
>>>
>>>
>>>
>>>
>>>
>>
>>
>
Other related posts: