HPE 3PAR StoreServ Storage

Cannot launch dismisspd 43 - Error pd id referenced by another chunklet

 
SOLVED
Go to solution
FS-EXPLOITATION
Occasional Contributor

Cannot launch dismisspd 43 - Error pd id referenced by another chunklet

Hi,

We have a failled disk which seem to be impossible  to dismiss.

The command dismisspd failled with an error message :

Error : Pd id 43 is referenced by chunklet 1:1584

We've also tried to lauch the command to check how many reference exist and we've found 1129 chunklets presents on other disks.

showpdch -mov

Here's a part of the results we're getting:

Pdid Chnk LdName LdCh State Usage Media Sp Cl From To
1 1584 tp-19-sd-0.6 136 normal ld valid Y N 43:77 ---
1 1585 tp-17-sd-0.0 254 normal ld valid Y N 43:122 ---
1 1586 tp-17-sd-0.2 78 normal ld valid Y N 43:151 ---
1 1587 tp-15-sd-0.22 313 normal ld valid Y N 43:190 ---
1 1588 tp-8-sd-0.29 182 normal ld valid Y N 43:216 ---
1 1589 tp-8-sd-0.6 318 normal ld valid Y N 43:38 ---
1 1590 tp-15-sd-3.0 282 normal ld valid Y N 43:106 ---
1 1591 tp-15-sd-0.67 326 normal ld valid Y N 43:225 ---
1 1592 tp-8-sd-0.31 189 normal ld valid Y N 43:265 ---
1 1593 tp-15-sd-0.38 158 normal ld valid Y N 43:290 ---
1 1594 tp-15-sd-0.30 204 normal ld valid Y N 43:311 ---

Is there a way to entirely deferenced all chunklet referring to the pid 43 ?

I've also tried the command to erase the chunklet but also i'm getting is an error:

movech -perm -ovrd -f -dr 1:1584 1:1584

Error: Disk 13 chunklet 1584 is not in use.

 

Thanks in advance for your insights,

OHC

1 REPLY 1
Anjum_IK
HPE Pro
Solution

Re: Cannot launch dismisspd 43 - Error pd id referenced by another chunklet

@FS-EXPLOITATION 

Good day!


This happens if the drive failed and then came back to life and system relocated some chunklets back to the drive. This could also happen if the servicemag resume command failed for some reason

The workaround is to relocate the chunklets that are referenced by the old PDID to valid locations without affecting the availability.

Also do check servicemag status. if it has failed, unmark, clear and re-run resume. Then after resume succeeds, dismiss the old id

OR

The command you can run to do this task manually assuming the servicemag has succeeded :

movech -perm -nowait 43:77 (for each chunklet)

1000+ chunklets is alot

There is script in the system which will move all the chunklets one by one automatically.

Request you to open the ticket with Hpe support. (To have the script work for you)

I work for HPE


If you feel this was helpful please click the KUDOS! thumb below!
Accept or Kudo


[Any personal opinions expressed are mine, and not official statements on behalf of Hewlett Packard Enterprise]