Re: acfsutils on 11.2.0.4 - Rename mountpoint - Solved
From: David Barbour <david.barbour1_at_gmail.com>
Date: Wed, 4 Nov 2015 13:37:15 -0600
Message-ID: <CAFH+ife6F+n5QMqbr7nA7x7bAAzhVfNR4bqkPiJ7jdynmnuwYQ_at_mail.gmail.com>
Okay - it does work. As long as you unmount the filesystem first! Interesting that acfsutils reports stuff with the 'info' modifier, but the registry modifier is ineffective.
ora.grid.t28base.acfs
/rchr2t01/
/root # srvctl remove filesystem -d /dev/asm/t28base-163
ora.grid.t28base.acfs
/rchr2t01/
/root # ls -al /oracle/TEST
total 76
Date: Wed, 4 Nov 2015 13:37:15 -0600
Message-ID: <CAFH+ife6F+n5QMqbr7nA7x7bAAzhVfNR4bqkPiJ7jdynmnuwYQ_at_mail.gmail.com>
Okay - it does work. As long as you unmount the filesystem first! Interesting that acfsutils reports stuff with the 'info' modifier, but the registry modifier is ineffective.
Thanks for the input.
Note that I have my ACFS filesystem on ASM disk. What worked was the following:
/rchr2t01/
/root # ls -al /oracle/TESTTemp
total 76
drwxrwxr-x 5 oracle oinstall 4096 Oct 29 15:36 . drwxr-xr-x 22 root oinstall 4096 Oct 29 09:53 .. drwxr-xr-x 11 oracle oinstall 4096 Oct 29 15:36 diag drwx------ 2 root root 65536 Oct 29 15:34 lost+found
/rchr2t01/
/root # ls -al /oracle/TEST
total 8
drwxr-xr-x 2 oracle oinstall 4096 Oct 29 09:52 .
drwxr-xr-x 22 root oinstall 4096 Oct 29 09:53 ..
/rchr2t01/
/root # srvctl config filesystem -d /dev/asm/t28base-163
Volume device: /dev/asm/t28base-163
Canonical volume device: /dev/asm/t28base-163
Mountpoint path: /oracle/TESTTemp
User: oracle
ACFS file system is enabled
/rchr2t01/
/root # srvctl stop filesystem -d /dev/asm/t28base-163
/root # crsctl stat res -t
NAME TARGET STATE SERVER STATE_DETAILS --------------------------------------------------------------------------------Local Resources
ora.grid.t28base.acfs
OFFLINE OFFLINE rchr2t01 admin
unmounted /or
acle/TESTTemp
OFFLINE OFFLINE rchr2t02 admin
unmounted /or
acle/TESTTemp
/rchr2t01/
/root # srvctl remove filesystem -d /dev/asm/t28base-163
/rchr2t01/
/root # srvctl add filesystem -d /dev/asm/t28base-163 -g GRID -v T28BASE -m
/oracle/TEST -u oracle
/rchr2t01/
/root # srvctl start filesystem -d /dev/asm/t28base-163
/rchr2t01/
/root # crsctl stat res -t
NAME TARGET STATE SERVER STATE_DETAILS --------------------------------------------------------------------------------Local Resources
ora.grid.t28base.acfs
ONLINE ONLINE rchr2t01 mounted on
/oracle/
TEST
ONLINE ONLINE rchr2t02 mounted on
/oracle/
TEST
/rchr2t01/
/root # ls -al /oracle/TEST
total 76
drwxrwxr-x 5 oracle oinstall 4096 Oct 29 15:36 . drwxr-xr-x 22 root oinstall 4096 Oct 29 09:53 .. drwxr-xr-x 11 oracle oinstall 4096 Oct 29 15:36 diag drwx------ 2 root root 65536 Oct 29 15:34 lost+found
On Wed, Oct 28, 2015 at 5:39 PM, Ls Cheng <exriscer_at_gmail.com> wrote:
> uh... I did this a while back in 11.2.0.4 and it worked, srvctl remove > filesystem should not remove the FS or volume > > On Wed, Oct 28, 2015 at 11:06 PM, David Barbour <david.barbour1_at_gmail.com> > wrote: > >> Hint: Removing and adding back via srvctl deletes the data in the >> filesystem. I know. It was sort of my command of last resort - but I did >> make a copy since I had a really bad feeling about running it. >> >> On Wed, Oct 28, 2015 at 3:23 PM, Ls Cheng <exriscer_at_gmail.com> wrote: >> >>> Hi >>> >>> In 11.2 srvctl modify filesystem hasnt go the -path option as 12cR1 so >>> the only supported way seems to be remove the file system from CRS and >>> added back using srvctl. The other way might be crsctl modify resource attr >>> MOUNTPOINT_PATH but it is unsupported >>> >>> >>> Thanks >>> >>> On Wed, Oct 28, 2015 at 8:40 PM, David Barbour <david.barbour1_at_gmail.com >>> > wrote: >>> >>>> Thanks. I'll try some different combinations and see if anything >>>> works. I did see your reference yesterday when nothing was working as >>>> expected and I resorted to asking my friend Google. As mentioned, I've got >>>> a number of these still left to do so I can do some prior testing and >>>> determine what might be different in my environment or what might work. >>>> >>>> On Wed, Oct 28, 2015 at 2:20 PM, viljo <hakala.viljo_at_gmail.com> wrote: >>>> >>>>> Maybe a good time to open an SR with MOS >>>>> as you are getting unexpected results? >>>>> >>>>> Should work as documented here: >>>>> >>>>> >>>>> >>>>> http://appcrawler.com/wordpress/2012/02/21/change-the-mount-point-for-an-acfs-filesystem/ >>>>> >>>>> My experience with ACFS is based on 10gr2 and >>>>> 11.2.0.4. It works if you dont do changes to the environment. Any kind >>>>> of changes can trigger numerous bugs. Last case for us was that we had to >>>>> shutdown ASM instance due to a bug that prevented storage migration/old >>>>> disks from being removed from acfs volume diskgroup. We ended up removing >>>>> acfs totally and getting a NFS share instead. Performance isn't as good but >>>>> it feels more stable. >>>>> >>>>> ALSO: >>>>> Recently ASM instances hanged nearly every time in version 11.2.0.4 >>>>> when we did storage migrations. Should be online add/drop disk operations, >>>>> but a combination of oracle vm + oel + 11.2.0.4 clusterware really has >>>>> proved to be nothing like you would expect in this century. The ASM hangs >>>>> are clusterwide. Databases can talk to ASM but disk group queries and >>>>> asmcmd just hang. Only entire cluster reboots solve the hangs. If one >>>>> rebooted only one node, that was used to run the add/drop disk commands, >>>>> OCR disk group on that node became inaccessible according to CRS alert.log, >>>>> but everything was OK when checked and othet nodes were responding to DG >>>>> queries. >>>>> >>>>> >>>>> David Barbour <david.barbour1_at_gmail.com> kirjoitti 28.10.2015 kello >>>>> 20.13: >>>>> >>>>> The ACFS is a cluster managed resource. Since I posted this I tried a >>>>> couple of different commands and ending up wiping out the D28Temp ACFS >>>>> filesystem and have since recreated it and mounted it where I wanted it, >>>>> but I need to go through this exercise for several other databases and >>>>> would appreciate input. >>>>> >>>>> acfsutil registry -l returns no values, although acfsutil info fs does >>>>> >>>>> /oracle/grid/11204/install/usm/cmds/bin # crsctl stat res -t >>>>> >>>>> -------------------------------------------------------------------------------- >>>>> NAME TARGET STATE SERVER >>>>> STATE_DETAILS >>>>> >>>>> -------------------------------------------------------------------------------- >>>>> Local Resources >>>>> >>>>> -------------------------------------------------------------------------------- >>>>> ora.grid.d28base.acfs >>>>> ONLINE ONLINE rchr2t01 mounted >>>>> on /oracle/D28 >>>>> ONLINE ONLINE rchr2t02 mounted >>>>> on /oracle/D28 >>>>> >>>>> /oracle/grid/11204/install/usm/cmds/bin # acfsutil info fs >>>>> /oracle/D04 >>>>> ACFS Version: 11.2.0.4.0 >>>>> flags: MountPoint,Available >>>>> mount time: Fri Sep 18 14:35:18 2015 >>>>> volumes: 1 >>>>> total size: 17179869184 >>>>> total free: 10023948288 >>>>> primary volume: /dev/asm/d04base-163 >>>>> label: >>>>> flags: Primary,Available,ADVM >>>>> on-disk version: 39.0 >>>>> allocation unit: 4096 >>>>> major, minor: 252, 83457 >>>>> size: 17179869184 >>>>> free: 10023948288 >>>>> ADVM diskgroup GRID >>>>> ADVM resize increment: 33554432 >>>>> ADVM redundancy: unprotected >>>>> ADVM stripe columns: 4 >>>>> ADVM stripe width: 131072 >>>>> compatible.advm: 11.2.0.0.0 >>>>> number of snapshots: 0 >>>>> snapshot space usage: 0 >>>>> replication status: DISABLED >>>>> >>>>> /oracle/T04 >>>>> ACFS Version: 11.2.0.4.0 >>>>> flags: MountPoint,Available >>>>> mount time: Mon Oct 26 15:59:15 2015 >>>>> volumes: 1 >>>>> total size: 17179869184 >>>>> total free: 11939848192 >>>>> primary volume: /dev/asm/t04base-163 >>>>> label: >>>>> flags: Primary,Available,ADVM >>>>> on-disk version: 39.0 >>>>> allocation unit: 4096 >>>>> major, minor: 252, 83458 >>>>> size: 17179869184 >>>>> free: 11939848192 >>>>> ADVM diskgroup GRID >>>>> ADVM resize increment: 33554432 >>>>> ADVM redundancy: unprotected >>>>> ADVM stripe columns: 4 >>>>> ADVM stripe width: 131072 >>>>> compatible.advm: 11.2.0.0.0 >>>>> number of snapshots: 0 >>>>> snapshot space usage: 0 >>>>> replication status: DISABLED >>>>> >>>>> /oracle/D28 >>>>> ACFS Version: 11.2.0.4.0 >>>>> flags: MountPoint,Available >>>>> mount time: Wed Oct 28 12:14:32 2015 >>>>> volumes: 1 >>>>> total size: 17179869184 >>>>> total free: 11243798528 >>>>> primary volume: /dev/asm/d28base-163 >>>>> label: >>>>> flags: Primary,Available,ADVM >>>>> on-disk version: 39.0 >>>>> allocation unit: 4096 >>>>> major, minor: 252, 83459 >>>>> size: 17179869184 >>>>> free: 11243798528 >>>>> ADVM diskgroup GRID >>>>> ADVM resize increment: 33554432 >>>>> ADVM redundancy: unprotected >>>>> ADVM stripe columns: 4 >>>>> ADVM stripe width: 131072 >>>>> compatible.advm: 11.2.0.0.0 >>>>> number of snapshots: 0 >>>>> snapshot space usage: 0 >>>>> replication status: DISABLED >>>>> >>>>> >>>>> On Wed, Oct 28, 2015 at 12:34 PM, Ls Cheng <exriscer_at_gmail.com> wrote: >>>>> >>>>>> Hi David >>>>>> >>>>>> What erros are you getting? >>>>>> >>>>>> Is your ACFS a general purpose FS or CRS managed (a resource under >>>>>> Clusterware control) >>>>>> >>>>>> May be you can share a crsctl stat res -t output and acfsutil >>>>>> registry -l output? >>>>>> >>>>>> Thanks >>>>>> >>>>>> >>>>>> On Wed, Oct 28, 2015 at 5:42 PM, David Barbour < >>>>>> david.barbour1_at_gmail.com> wrote: >>>>>> >>>>>>> Has anyone renamed an acfs mount point using acfsutils in 11.2.0.4? >>>>>>> >>>>>>> Running GI 11.2.0.4 and upgrading the databases to 11.2.0.4 from >>>>>>> 11.2.0.3. Due to our naming conventions, I installed the 11.2.0.4 binaries >>>>>>> on a new ASM Cluster File System which I mounted on /oracle/<SID>Temp. Now >>>>>>> that everything is working, I want to mount this on /oracle/<SID>. >>>>>>> >>>>>>> In the past I was able to use acfsutil to perform this type of >>>>>>> action, but I'm getting unexpected results using acfsutil registry >>>>>>> commands. >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >
-- http://www.freelists.org/webpage/oracle-lReceived on Wed Nov 04 2015 - 20:37:15 CET