Delete AO configuration and return the data to original FC tier;

[Carlos_Oliveira]

Hello 3PAR team,

I need your opinion, my customer  has FC and SSD disks in 3PAR8200 with AO configured ... SSD disks are exclusive only for AO and nothing else ... the fact is that now the customer wants to permanently delete the AO configuration and use SSDs for other purposes! However, due to AO, there is data on the SSD, what is the procedure, command line, to return the data that is on the SSD to the FC source disks?

HPE 3PAR 8200 version 3.3.1.410 (MU2)

Regards,

Carlos Oliveira.

[sbhat09]

Hello @Carlos_Oliveira,

Sure. I will help you on that. Can you please share the below command outputs to provide you precise instructions? this

showaocfg
showcpg

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.

[Carlos_Oliveira]

Hello Srinivas Bhat, first of all thank you very much for the quick response ... follow the output of the commands .....

[sbhat09]

Hello @Carlos_Oliveira,

Thanks for the details. Here the free space in the FC tier is more than the used space of SSD by the AO. So, you are good to move all the data from SSD back to the VV's parent CPGs (FC).

To offload the SSD space, used by AO, set the SSD CPG warning level to 1 MB. (login to SSMC -> main menu -> CPGs -> select the cpg 'AO_CPG_R6_SSD_VMWARE' -> Actions -> edit -> enable growth warning -> set it to 1 MB. Repeat the same for 'AO_CPG_R6_SSD_ORACLE' as well). Leave it until the next 1 or 2 AO schedules run. You will see the entire SSD space drained out. Then you can detach the SSD CPG from the AO or remove the AO configuration completely.

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.

[Carlos_Oliveira]

Hello Srinivas Bhat, many thanks for your help, yesterday I sent the procedure to the customer, because I had no access to the storage, the customer ran it without problems and we will monitor until tomorrow to see if the SSD disks empty ....

But the customer also accidentally performed the procedure in another AO CPG ... ... is there any problem doing this procedure in another AO CPG that was not SSD ? How to revert ? It was like this :

AO_CPG_R6_SSD_VMWARE --- Done and monitoring;
AO_CPG_R6_SSD_ORACLE --- Done and monitoring;
AO_CPG_R6_FC_VMWARE --- This done wrong! Have a problem ? How to revert to the original state before the change?

Regards,

Carlos Oliveira.

[sbhat09]

Hello @Carlos_Oliveira,

That will not do much harm. Customer would have received an alert for space warning. Thats it.

That operation will just push the other CPG data back to the parent CPGs. They can disable that warning for the wrong CPG now. No worries.

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.

[sbhat09]

Hello @Carlos_Oliveira,

Just curious to know how it all went. Do you need any more assistance?

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.

[Carlos_Oliveira]

Hello Srinivas Bhat, how are you ?

Thank you for your usual contact ...

It seems to me that one CPG_ORACLE is ok and empty... but the other one CPG_VMWARE doesn't empty and I've performed the procedure for days... Do you have any idea or tip of what it could be? Here are screenshots taken today:

[sbhat09]

Hello @Carlos_Oliveira

Data is not offloading from the 'AO_CPG_R6_SSD_VMWARE' because, in the same AO configuration, you have also set the 1MB warning limit to the 'AO_CPG_R6_FC_VMWARE'. So, data doesn't move eitherways.
Disable the 1MB warning limit to 'AO_CPG_R6_FC_VMWARE' and let the next AO schedule to run. You will see the expected data movement back to the 'AO_CPG_R6_FC_VMWARE'.

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.

[sbhat09]

Hello @Carlos_Oliveira,

Hope you are able to achieve the expected data movement back to the FC CPG now. (If you had disabled the 1MB warning limit for 'AO_CPG_R6_FC_VMWARE')

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.

[Carlos_Oliveira · Accepted solution]

Hello , @Srinivas Bhat , 

Thank you very much for your support and follow-up, all the procedures worked perfectly, all the data from the CPG on SSD made the proper migration to the origin and we were able to meet the customer's expectation... t

Thank you very much.

Regards,

Carlos Oliveira.

[sbhat09]

Hello @Carlos_Oliveira,

That is great to know from you. Thanks for updating.!

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.

[Jonnycoel]

Hi Sbhat, that is excellent trick. thank you. I am also facing similar proiblem. We are making some changes. We want to remove FC raid 1 cpg from AO configuration and add FC raid 6 cpg. Can I follow the same steps you advised?

[sbhat09]

Hello @Jonnycoel,

Yes. You can follow the below steps to drain the FC RAID1 cpg. When the CPG utilization is completely nil, you will be able to remove the CPG from the AO configuration. After that you can add FC RAID6 cpg to the AO configuration.

"To offload the SSD space, used by AO, set the SSD CPG warning level to 1 MB. (login to SSMC -> main menu -> CPGs -> select the cpg 'AO_CPG_R6_SSD_VMWARE' -> Actions -> edit -> enable growth warning -> set it to 1 MB. Repeat the same for 'AO_CPG_R6_SSD_ORACLE' as well). Leave it until the next 1 or 2 AO schedules run. You will see the entire SSD space drained out. Then you can detach the SSD CPG from the AO or remove the AO configuration completely."

Please let me know if you need further assistance.

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.

[Jonnycoel]

Ok. Thanks you. I will try this.

[sbhat09]

Hello @Jonnycoel,

Have you tried it? Did that fix the issue?

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.

[Jonnycoel]

That worked like wonder. Thank you so much.!

[Sunitha_Mod]

Hello @Jonnycoel,

That's Awesome! 

We are glad to know the problem had been resolved.