This post is more than 5 years old
51 Posts
0
12614
replace meta volumes
Hello,
we are replacing the array behind the vplex so we also have to change the meta volumes on it.
Do you know what is the best way to do it?
Regards
This post is more than 5 years old
51 Posts
0
12614
Hello,
we are replacing the array behind the vplex so we also have to change the meta volumes on it.
Do you know what is the best way to do it?
Regards
Top
tdaramsay
16 Posts
0
March 12th, 2015 17:00
Just bumping this as i have to currently decomission an array and move the meta & backup volumes.
The procedure I was given was to
1) Create a new meta-volume from two disks that are meta-volume candidates
2) use the meta-volume move command to move the current active volume to the new volume
3) remove the old. unactive meta-volume.
How did you're procedure go?
Aglidic
51 Posts
0
February 2nd, 2015 01:00
i was thinking to use the detach-mirror and attach-mirror command to do that
Intech1
57 Posts
0
February 4th, 2015 08:00
Hi
meta-volume attach-mirror --storage volume "vpd-of-meta-copy-to-attach" --meta-volume "your meta volume"
Once in sync:
Monitor with ll /path-to-your-meta look for rebuild-progress, rebuild-status, rebuild-eta
Then do:
meta-volume detach-mirror --storage-volume "vpd-of-meta-copy-to-detach" --meta-volume "your meta volume"
Steve
Sent from my iPad
TadhgConcannon
1 Rookie
1 Rookie
•
22 Posts
0
February 10th, 2015 04:00
Hi all
You can also set the speed of rebuild from from the default 128K to 128M -- from within the meta volume
set transfer-size 128M
/clusters/Cluster-2/system-volumes/SiteB_VPLEX_Meta_1
Also remember you need to add 2 x new Meta Volumes of 78GB
You also need to move the backup volumes to a temporary location ( 2 Backup volumes required)
Same for the logging volumes (2 x logging volumes required)
Once that is complete then remove the old array
Follow the procedure generator to remove the old aray
I'm doing this at the moment removing an old VNX and replacing with a new VNX. Any questions please let me know
TadhgConcannon
1 Rookie
1 Rookie
•
22 Posts
2
March 13th, 2015 02:00
Hi all
some screen shots on the procedure I have roughtly documented
Meta Volume Backup
cd to
/clusters/Cluster-2-XXX/system-volumes
run
configuration show-meta-volume-candidates
This should show you volumes whihch are 78GB in size and are suitable candidates for Meta Volume backup
run
configuration metadata-backup
then select the required volumes for Meta Volumes to change the Meta volume disks to the new disks
Meta-Volume Detach Mirror
cd to
VPlexcli:/clusters/Cluster-2-XXX/system-volumes/SiteB_VPLEX_Meta_1/components
to remove the META VOLUMES
meta-volume detach-mirror --storage-volume VPD83T3:MeaningfulText --meta-volume SiteB_VPL EX_Meta_1
meta-volume detach-mirror --storage-volume VPD83T3:Meaningful text--meta-volume SiteB_VPL EX_Meta_1
Meta Volumes removed
Moving a logging Volumes
cd to
/clusters/Cluster-2-XXX/system-volumes/SiteB-VPLEX-Log_vol/components
run ls -la
Name Slot Type Operational Health Capacity
------------------------------------ Number ------ Status State --------
------------------------------------ ------ ------ ----------- ------ --------
extent_VPLEXNAME_SiteB-VPLEX-Log_1_1 0 extent ok ok 10G
extent_VPLEXNAME_SiteB-VPLEX-Log_2_1 1 extent ok ok 10G
Above are the current logging volumes
There are two logging volumes
Create NEW Logging Volumes from extents
logging-volume create -n SiteB-VPLEX-XIO-log_vol -g raid-1 -e extent_XIO_SiteB_SiteB_VPLEX_Log_1_1, extent_XIO_SiteB_SiteB_VPLEX_Log_2_1
cd to
cd /distributed-storage/distributed-devices/
Set Logging Volumes ll
set-log --logging-volumes SiteB-VPLEX-XIO-log_vol_vol --distributed-device * (type in do not copy and include the *)
check this worked
cd to
/clusters/Cluster-2-VPLEXNAME/system-volumes/SiteB-VPLEX-log_vol_vol/segments>
then run ls -la
check original logging volume is not in use by any distributed devices
cd to
/clusters/Cluster-2-VPLEXNAME/system-volumes/SiteB-VPLEX-Log_vol/segments>
ls -la
Name Starting Block Block Count Use
------ -------------- ----------- ----
free-0 0 5440 free
Destroy the logging volume
logging-volume destroy --logging-volume SiteB-VPLEX-Log_vol
Destroy the extents behind the logging volume
extent destroy --extents extent_VNX/VMAX_SiteB-VPLEX-Log_1_1
extent destroy --extents extent_VNX/VMAX_SiteB-VPLEX-Log_1_1
unclaim the volume from storage
check out the VPD / storage volume name this needs to be unclaimed
VPlexcli:/clusters/Cluster-2-VPLEXNAME/storage-elements/storage-volumes> unclaim VNX_SiteB_SiteB-VPLEX-Log_1
Unclaimed 1 of 1 storage-volumes.
VPlexcli:/clusters/Cluster-2-VPLEXNAME/storage-elements/storage-volumes> unclaim VNX_SiteB_SiteB-VPLEX-Log_2
Unclaimed 1 of 1 storage-volumes.
Hope this helps
Regards
Tadhg
TadhgConcannon
1 Rookie
1 Rookie
•
22 Posts
0
March 23rd, 2015 02:00
Hi Andy
Good to hear it worked
Regards,
Tadhg
tdaramsay
16 Posts
1
March 23rd, 2015 02:00
Thanks for the update. I tried your method and it worked perfectly!
Menno
1 Rookie
1 Rookie
•
29 Posts
0
January 25th, 2018 02:00
Andy,
i have five 10gb log volumes. i want to move them to new array's. created 5 new 10gb logging volumes. but move will not work.
set-log --logging-volumes new_loggingvol_1 --distributed-devices *
the error i get:
couldn't find space on a healthy reserved logging device.
any idea? thanks in advance!
mwright2
11 Posts
1
May 9th, 2018 13:00
Since I recently performed this exact activity, I am providing the steps i used.
Meta Volume Moves to new array
present 4 new 80gb volumes from new array - DO NOT CLAIM THEM thru VPLEX
navigate to /clusters/cluster-1/system-volumes
run ll command - to list current system volumes
should show the following
cluster1 logging volume
cluster1 meta volume
cluster1 meta volume backup1
cluster1 meta volume backup2
navigate to /clusters/cluster-1/system-volumes/Cluster1-Meta-Volume
cd /components
run ll - should list the 2 VPD's of the 2 current volumes(luns) used for the meta volume
run command - configuration show-meta-volume-candidates
- you should see the 4 new 80gb volumes from the new array with VPD's
cd /clusters/cluster-1/system-volumes/Cluster1-Meta-Volume
run command - meta-volume detach-mirror -d VPD83T3:6008888888888d88bef88888888de888 --meta-volume CL1_META - this should be the VPD# from one of the original 80gb luns which you want to remove
*** attach new volume from new array
run command - meta-volume attach-mirror -d VPD83T3:600777777777777bef777777777de777 --meta-volume CL1_META - this should be the VPD# from one of the new 80gb luns from the new array
cd /clusters/cluster-1/system-volumes/Cluster1-Meta-Volume/components
run ll
newly added mirror will show error and critical failure - no worries - just need to wait while it syncs up
if you navigate to the /clusters/cluster-1/system-volumes/Cluster1-Meta-Volume/ and run the "ll" command it will provide an rebuild-ETA and percentage complete
wait until the rebuild process completes
navigate to /clusters/cluster-1/system-volumes
run the "ll" command
verify "health state" is ok for all 4 volumes
navigate to /clusters/cluster-1/system-volumes/Cluster1-Meta-Volume
cd /components
run ll - should list the 2 VPD's - one from each array now
repeat the meta-volume detach removing the old array
repeat meta-volume attach for the second leg of the cluster1 meta-volume using the new array VPD
wait for sync to complete
navigate to /clusters/cluster-1/system-volumes
verify Operational Status and Health Status of all volumes
navigate to /clusters/cluster-1/system-volumes/cluster1-metavolume/components
run ll
verify both VPD's are from new array
navigate to /clusters/cluster-1/system-volumes
- we still need to delete the 2 backup volumes and add the new ones from the new array
cd to one the MetaVolume backups
cd to the components - verify VPD is from old array
***delete the backup
run command - meta-volume destroy --meta-volume
run meta volume destroy command twice - one for each backup
navigate to /clusters/cluster-1/system-volumes
run ll - verify logging and Meta volume exist ( should be 2 in the output)
verify you still have 2 80gb volumes unclaimed to use for meta volumes backup
run command - configuration show-meta-volume-candidates
Reconfigure the backup schedule - this will allow you to use the 2 new volumes from the new array
cd /clusters/cluster-1/system-volumes
run command - configuraton metadata-backup
When asked "Do you want change the existing schedule?" - Select Y
It should list the 2 new volumes from the new array
you will have to specify the 2 volumes to use for backups
Please select volumes for meta-data backup, preferably from two different arrays (volume1,volume2):VPD#1,VPD#2
VPLEX is configured to back up meta-data every day at 23:00 (UTC).
Would you like to change the time the meta-data is backed up? [no]: You can select NO here if you don't want to the change the time
It should specify a summary with the backup time and which volumes it will use - verify VPD's are from new array
Would you like to run the setup process now? [yes]: yes - select YES for this question
Scheduling the backup of metadata...
Performing metadata backup (This will take a few minutes) - it will perform a backup which we can verify later
Successfully performed the initial backing up of metadata
Successfully scheduled the backing up of metadata
Successfully scheduled the metadata backup
The metadata backup has been successfully scheduled.
navigate to /clusters/cluster-1/system-volumes
run ll
you should now see 4 volumes listed
Logging
MetaVolume
MetaVolume1 - backup
MetaVolume2 - backup
You can change to the components folder of the MetaVolume, MetaVolume 1 Backup, and MetaVolume 2 backup to verify the VPD's are from the new array
***you can also verify the consistency on the meta volumes
run command - meta-volume verify-on-disk-consistency -c cluster
Unclaim 4 old luns
unmap from storage array
rediscover array and "forget unreachable luns"
run health-check or health-check --full