3Par 7200 remove disk

[Tecniciz]

Hi to all,
i need to remove disk from a cage of 3Par 7200 and normally i use this procedure:

setpd ldalloc off <DiskID>
movepdtospare -f -vacate -nowait <DiskID>
showpdch -spr <DiskID>      
removespare PDID:a           
dismisspd <DiskID>                                                      

usually this procedure works but now I have a problem with about 10 disks, take the disk with ID 163 as an example

showpd -s
1163 7:6:0 FC degraded not_available_for_allocations not_capable

setpd ldalloc off 161
showpd -s

163 7:6:0 FC degraded not_available_for_allocations not_capable

cli% movepdtospare -f -vacate -nowait 163
Error: Could not find relocation target for one or more chunklets

cli% showpdch -mov | grep 163
No chunklet information available.

cli% showpdch -spr 163
No chunklet information available.

cli% dismisspd 163
Error : Pd id 163 is in use

I don't understand why the commands on these disks don't work, can you give me some advice?

Thank you
Angelo

[sbhat09]

Hello @Tecniciz,

At first, please check if there is enough spare free and spare chunkles available to move the chunklets from the drive.
'showpd -c' command output will help you to gain some better insights.
Based on that information, you can try to dig more into the issue like which PDs have problems moving chunklets, which LDs, which are the exact problematic chunklets.

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Tecniciz]

Hi Srinivas,
thx in advanced for your support
Here below the output of showpd -c command referring to the disks in cage 7 that I have to free:

Are in degrade state because i have set the command setpd ldalloc off <DiskID>

I need some more advice on how to solve my problem, if you have commands to run for example.

Thank you
Angelo

[sbhat09]

You are most welcome @Tecniciz.!

When you had run movepdtospare -f -vacate -nowait <DiskID> which is the disk id was specified as relocation target? Do you see enough free/spare chunklets in the relocation target as per 'showpd -c' output?

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Tecniciz]

Hi Srinivas,
when i use the command movepdtospare -f -vacate -nowait <DiskID> the diskID i set the disk that i want remove from 3Par.

For example if i want migrate all chunklets from disk with ID 163 because i want remove this disk from 3Par i set the command:
movepdtospare -f -vacate -nowait 163

About you second question:
Do you see enough free/spare chunklets in the relocation target as per 'showpd -c' output?

I don't understand, what should I look for in the showpd -c command output, can you give me directions?

Regards
Angelo

[sbhat09]

Hello @Tecniciz,

Just run the 'showpd -c'  command output and check if you have sufficient free and spare chunklets in rest of the drives of same type. (Same type as the drives you are removing).

Just to understand if the rest of the drives have enough free space to accommodate the moving chunklets.

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Tecniciz]

Here below the details of command showpd -c

In my opinion we have enough free space for chunklet movement

If you can let me know please

Angelo

[sbhat09]

Hello @Tecniciz,

Sorry, I was caught up with my work and couldn't respond to you.

If you still have the issue, it could be due to the failed chunklet in disk 163. You can identify the chunklet by 'showpdch -fail -from 163'
Then move the chunklet to any other disk which is part of the same CPG.
movech <disk_id>:<chunklet_position_id>-<target_disk_id>:<target_chunklet_position_id>
Example to move 49th chunklet from disk 163 to 500th position of disk 144 'movech 163:49-144:500'

If you see further issues, you can refer this dicussion chain where I had provided similar inputs - https://community.hpe.com/t5/hpe-3par-storeserv-storage/3par-7200c-disk-issue/m-p/7150430#M7010 

Also, I could observe poor chunked balancing across the drives. So, consider to run tunesys operation at cpg levels for every CPGs (If you run it at system level, it may take very long time at a stretch). The command is (tunesys -cpg <cpg_name>)

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Tecniciz]

Hi Srinivas,
thx for your replay.

Now i have another problem, i have I run the movepdtospare PID command for disk with ID 149, 150, 151 and 152 but the migration of chunklets is blocked, chunklets don't have destination disks, below a screen.
Is it possible to block this transfer job? if i run the showtask command there is no active task.

if you can let me know
Regards

Angelo

[sbhat09]

Hello @Tecniciz,

That could be because you are trying to move the chunklets from multiple PDs at once. Please try one at a time.

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Tecniciz]

@sbhat09 

Hi Srinivas ,
since the move of chunklets is blocked it does not advance, how can I interrupt it manually and start again doing one at a time?

[Tecniciz]

Hi Srinivas ,
it's possible interrupt it manually this move of chunklets and start again?

Let me know
Regards

Angelo

[Tecniciz]

Help

[sbhat09]

Hello @Tecniciz,

Please check if you find this discussion chain useful - https://community.hpe.com/t5/hpe-3par-storeserv-storage/3par-7200c-disk-issue/m-p/7150265#M7004

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.