Re: ASM disk to physical disk mapping
From: Vladimir M. Zakharychev <vladimir.zakharychev_at_gmail.com>
Date: Thu, 2 Jul 2009 10:39:08 -0700 (PDT)
Message-ID: <08aa9ff8-44cf-4c27-bb47-c8f27992908c_at_l31g2000yqb.googlegroups.com>
On Jul 2, 1:34 am, Dan <daniel.oster..._at_visaer.com> wrote:
> On Jul 1, 3:03 pm, John Hurley <johnbhur..._at_sbcglobal.net> wrote:
>
>
>
> > On Jul 1, 1:16 pm, Dan <daniel.oster..._at_visaer.com> wrote:
>
> > snip
>
> > > I checked the views and they give only what I can see in OEM.
> > > Unfortunately I don't have the create diskgroup statements because I
> > > stamped and added them to the diskgroup via asmtoolg, a gui-like tool
> > > that does all the work for you (another reason to do things
> > > yourself). I'm still researching this, I'll let you know what I
> > > find. I've seen others that lament this difficulty too, so I guess
> > > it's a known problem.
>
> > Don't forget that asm has an alert log file and ( well depending on
> > how long ago they occurred ) many important things that occur to an
> > asm instance ( like creating / mounting / etc diskgroups ) will show
> > up in that alert log.
>
> > ...
>
> > SQL> alter diskgroup prod_dg1 mount
> > WARNING: Deprecated privilege SYSDBA for command 'ALTER DISKGROUP
> > MOUNT'
> > NOTE: cache registered group PROD_DG1 number=1 incarn=0xd954fd6e
> > NOTE: cache began mount (first) of group PROD_DG1 number=1
> > incarn=0xd954fd6e
> > WARNING::ASMLIB library not found. See trace file for details.
> > NOTE: Assigning number (1,1) to disk (/storage/prod_dg1/disk2)
> > NOTE: Assigning number (1,2) to disk (/storage/prod_dg1/disk3)
> > NOTE: Assigning number (1,3) to disk (/storage/prod_dg1/disk4)
> > NOTE: Assigning number (1,0) to disk (/storage/prod_dg1/disk1)
> > NOTE: start heartbeating (grp 1)
> > kfdp_query(): 3
> > kfdp_queryBg(): 3
> > NOTE: cache opening disk 0 of grp 1: PROD_DG1_0000 path:/storage/
> > prod_dg1/disk1
> > NOTE: F1X0 found on disk 0 fcn 0.0
> > NOTE: cache opening disk 1 of grp 1: PROD_DG1_0001 path:/storage/
> > prod_dg1/disk2
> > NOTE: cache opening disk 2 of grp 1: PROD_DG1_0002 path:/storage/
> > prod_dg1/disk3
> > NOTE: cache opening disk 3 of grp 1: PROD_DG1_0003 path:/storage/
> > prod_dg1/disk4
> > NOTE: cache mounting (first) group 1/0xD954FD6E (PROD_DG1)
> > NOTE: cache recovered group 1 to fcn 0.129503
> > NOTE: LGWR attempting to mount thread 1 for diskgroup 1
> > NOTE: LGWR mounted thread 1 for disk group 1
> > NOTE: opening chunk 1 at fcn 0.129503 ABA
> > NOTE: seq=51 blk=6183
> > NOTE: cache mounting group 1/0xD954FD6E (PROD_DG1) succeeded
> > NOTE: cache ending mount (success) of group PROD_DG1 number=1
> > incarn=0xd954fd6e
> > kfdp_query(): 4
> > kfdp_queryBg(): 4
> > NOTE: Instance updated compatible.asm to 10.1.0.0.0 for grp 1
> > SUCCESS: diskgroup PROD_DG1 was mounted
> > SUCCESS: alter diskgroup prod_dg1 mount
>
> > ...
>
> > My output may look a little funny as we are running EMC powerpath on
> > linux and it generates some strange names for devices ( /dev/
> > emcpower* ) ... I use symbolic links from a directory structure /
> > storage/asm_diskgroup_name to point to the emc power path devices to
> > make it more clear what is being used where.
>
> That was a good suggestion which I hadn't thought of. I just checked
> though and it has the same configuration labels that I get in the
> other utilities. Somehow I need to associate the disk with the LUN
> number and Target ID. This is so low level that it doesn't seem to
> have a link.
>
> I guess what I could do is go up to the array and pull the bad disk
> out. I had wanted to logically remove the disk from ASM, but if I
> simulate a single disk failure by just yanking it out, it should be
> okay, right? The ASM will remove the disk automatically.
WHERE G.GROUP_NUMBER=D.GROUP_NUMBER
AND G.NAME='YOUR_DISK_GROUP_NAME'
Date: Thu, 2 Jul 2009 10:39:08 -0700 (PDT)
Message-ID: <08aa9ff8-44cf-4c27-bb47-c8f27992908c_at_l31g2000yqb.googlegroups.com>
On Jul 2, 1:34 am, Dan <daniel.oster..._at_visaer.com> wrote:
> On Jul 1, 3:03 pm, John Hurley <johnbhur..._at_sbcglobal.net> wrote:
>
>
>
> > On Jul 1, 1:16 pm, Dan <daniel.oster..._at_visaer.com> wrote:
>
> > snip
>
> > > I checked the views and they give only what I can see in OEM.
> > > Unfortunately I don't have the create diskgroup statements because I
> > > stamped and added them to the diskgroup via asmtoolg, a gui-like tool
> > > that does all the work for you (another reason to do things
> > > yourself). I'm still researching this, I'll let you know what I
> > > find. I've seen others that lament this difficulty too, so I guess
> > > it's a known problem.
>
> > Don't forget that asm has an alert log file and ( well depending on
> > how long ago they occurred ) many important things that occur to an
> > asm instance ( like creating / mounting / etc diskgroups ) will show
> > up in that alert log.
>
> > ...
>
> > SQL> alter diskgroup prod_dg1 mount
> > WARNING: Deprecated privilege SYSDBA for command 'ALTER DISKGROUP
> > MOUNT'
> > NOTE: cache registered group PROD_DG1 number=1 incarn=0xd954fd6e
> > NOTE: cache began mount (first) of group PROD_DG1 number=1
> > incarn=0xd954fd6e
> > WARNING::ASMLIB library not found. See trace file for details.
> > NOTE: Assigning number (1,1) to disk (/storage/prod_dg1/disk2)
> > NOTE: Assigning number (1,2) to disk (/storage/prod_dg1/disk3)
> > NOTE: Assigning number (1,3) to disk (/storage/prod_dg1/disk4)
> > NOTE: Assigning number (1,0) to disk (/storage/prod_dg1/disk1)
> > NOTE: start heartbeating (grp 1)
> > kfdp_query(): 3
> > kfdp_queryBg(): 3
> > NOTE: cache opening disk 0 of grp 1: PROD_DG1_0000 path:/storage/
> > prod_dg1/disk1
> > NOTE: F1X0 found on disk 0 fcn 0.0
> > NOTE: cache opening disk 1 of grp 1: PROD_DG1_0001 path:/storage/
> > prod_dg1/disk2
> > NOTE: cache opening disk 2 of grp 1: PROD_DG1_0002 path:/storage/
> > prod_dg1/disk3
> > NOTE: cache opening disk 3 of grp 1: PROD_DG1_0003 path:/storage/
> > prod_dg1/disk4
> > NOTE: cache mounting (first) group 1/0xD954FD6E (PROD_DG1)
> > NOTE: cache recovered group 1 to fcn 0.129503
> > NOTE: LGWR attempting to mount thread 1 for diskgroup 1
> > NOTE: LGWR mounted thread 1 for disk group 1
> > NOTE: opening chunk 1 at fcn 0.129503 ABA
> > NOTE: seq=51 blk=6183
> > NOTE: cache mounting group 1/0xD954FD6E (PROD_DG1) succeeded
> > NOTE: cache ending mount (success) of group PROD_DG1 number=1
> > incarn=0xd954fd6e
> > kfdp_query(): 4
> > kfdp_queryBg(): 4
> > NOTE: Instance updated compatible.asm to 10.1.0.0.0 for grp 1
> > SUCCESS: diskgroup PROD_DG1 was mounted
> > SUCCESS: alter diskgroup prod_dg1 mount
>
> > ...
>
> > My output may look a little funny as we are running EMC powerpath on
> > linux and it generates some strange names for devices ( /dev/
> > emcpower* ) ... I use symbolic links from a directory structure /
> > storage/asm_diskgroup_name to point to the emc power path devices to
> > make it more clear what is being used where.
>
> That was a good suggestion which I hadn't thought of. I just checked
> though and it has the same configuration labels that I get in the
> other utilities. Somehow I need to associate the disk with the LUN
> number and Target ID. This is so low level that it doesn't seem to
> have a link.
>
> I guess what I could do is go up to the array and pull the bad disk
> out. I had wanted to logically remove the disk from ASM, but if I
> simulate a single disk failure by just yanking it out, it should be
> okay, right? The ASM will remove the disk automatically.
Depends on ASM version and patchset level: some early releases didn't handle such failures all too well, so I'd do this as a last resort. What does
SELECT
disk_number
,mount_status
,header_status
,mode_status,
,state
,label -- disk label (don't confuse with ASM disk name)
,path -- OS path to the disk - what you are looking for
,udid -- Universal Device ID of the disk - might help, too
FROM V$ASM_DISK, D, V$ASM_DISKGROUP G
WHERE G.GROUP_NUMBER=D.GROUP_NUMBER
AND G.NAME='YOUR_DISK_GROUP_NAME'
-- optionally restrict the query to show only -- currently mounted members of the DG -- AND (D.MOUNT_STATUS='CACHED' OR D.MOUNT_STATUS='OPENED') -- AND D.HEADER_STATUS='MEMBER'
show when executed in the ASM instance? Never dealt with ASM on Windows, so not sure if this will give you the information you need, but worth a try anyway. Note that you need to run this in ASM instance, not in the Oracle instance connected to this ASM instance.
Hth,
Vladimir M. Zakharychev
N-Networks, makers of Dynamic PSP(tm)
http://www.dynamicpsp.com
Received on Thu Jul 02 2009 - 12:39:08 CDT