vmax- R2 size not same as R1

Question

Hello All,

I need some help with a script that would generate cylinder sizes of all the R1 -R2 pair in the rdf group so that I could identify weather there is a mismatch in cylinder sizes . I noticed few mismatch ( R2 cylinders larger that R1) after I encountered an issue while running establish after the split was done for application testing . Support asked me to run -full est as a workaround .

So I need to identify weather there is any such devices in the entire environment . Also, after I identify the mismatch , how do I fix it?

Experts: Kindly assist .

Its VMAX 40K - 5876

Thanks !

umichklewis · Answer

You can do this from SYMCLI very easily. Use "symdev -sid XXXX -cyl" as below:

U:\>symdev -sid 1234 list -cyl

Symmetrix ID: 000199701234

Device Name Directors Device

--------------------------- ------------- -------------------------------------

Cap

Sym Physical SA : P DA :IT Config Attribute Sts (CYL)

--------------------------- ------------- -------------------------------------

00A4 Not Visible ***:* 07A:C0 2-Way Mir N/Grp'd ACLX WD 3

0355 Not Visible ***:* 08A:C1 2-Way Mir N/Grp'd RW 3

03AD Not Visible ???:? NA:NA TDEV N/Grp'd RW 109227

03AE Not Visible ???:? NA:NA TDEV N/Grp'd RW 109227

03AF Not Visible ***:* NA:NA TDEV Grp'd (M) RW 436912

You can easily use "grep" on Linux or "findstr" on Windows to find the device IDs. Or, you can try something like listing all the R1s with symrdf -sid XXXX list -R1 to get a list of your R1s, the pipe the output and run the devices through of symdev -sid XXXX show , grep'ing the cylinder counts.

Let us know if that helps!

Karl

tiro1 · Answer

I'd recommend you to check the devices size server by server, but you may need to do a little homework here. There will definitely be better ideas!

To check the devices size server by server

# symdev -sid list -sg -cyl | grep TDEV | awk '{print $1, $(NF)}' > tmp.txt ==> To get the size of the devices associated with the storage group.

# symdev -sid list -devs 2150:215D -v | egrep 'Device Symmetrix Name|Remote Device Symmetrix Name|Remote Symmetrix ID' ==> To get the partner devices devices ID and SID information.

Run the # symdev -sid list -sg -cyl command on target array with the appropriate SG name. Convert the data into a spreadsheet and compare.

I am 100% sure there will be other options too. Looking forward to see them.

deepsan_215421 · Answer

Thanks for the response .I could find the device , however , I need to fix this since we do DR testing and after failing over the RDF group it doesn't let me failback from a larger R2 to smaller R1 .

Please advise . Thanks!

tiro1 · Answer

One possible workaround would be delete the pairs which are in uneven size and create new R2 devices [of same size and configuration] and create the SRDF pair in the acp_disk mode. Obviously, you will have to notify app owners and liaise with platform teams to make changes in the DR servers. If the uneven R2s have clone/snaps, then you will have to reconfigure them completely.

deepsan_215421 · Answer

Thank you tyro & Karl for all the help .

The luns are presented to ESX hosts . I have RDF grp created and consistency is enabled . But , I am not clear of the steps . should I Suspend the RDF group first , then Unmount Volumes on DR host Unmap & Remap RDM in ESX , Re-mount Volumes on DR hosts and resume the replication ? How do I add the device pair back to the RDF group ..Please assist me by providing the steps:

Thanks

tiro1 · Answer

I'd highly recommend you to double-check this approach with EMC support before you make any changes to your production environment.

Step 1 - Unmount the uneven R2 RDMs from the DR ESXi hosts.

Step 2 - Remove the R2 devices from the Storage Groups and Unmap them from FAs.

Step 3 - Delete the SRDF relationship [You may have to create device file with the uneven R1/R2s, suspend the replication with cons_exempt option, and run deletepair]

Step 4 - Create new R2s [of correct size and configuration] and create SRDF relationship.

######Steps to add devices to SRDF/A RDFG

1. Create new R2s to match the R1 devices and prepare them to create SRDF pair.

2. Create SRDF pair. If you have any unused RDFG, please use it for this purpose. If you do not have, create one for this remediation work. Associate this temporary RDFG with as many RAs as possible - for better bandwidth.

#symrdf -sid <###> -f -rdfg createpair -type R1 -invalidate R2 -establish -rdf_mode acp_disk -nop

3. Monitor the copy progress

#symrdf -sid <###> -f -rdfg query

4. Once the devices are fully synchronized, suspend the replication.

#symrdf -sid <###> -f -rdfg verify -Synchronized

#symrdf -sid <###> -f -rdfg suspend -nop

5. Set the mode to async

#symrdf -sid <###> -f -rdfg set mode async -nop [assuming there are not too many invalid tracks to be copied]

6. Move the devices from the temporary RDFG to the appropriate RDFG [where you have the uneven R1/R2 devices]

#symrdf -sid <###> -f -rdfg movepair -new_rdfg -cons_exempt -nop

7. Resume the replication

#symrdf -sid <###> -f -rdfg resume -nop

8. Verify the status of the replication

#symrdf -sid <###> -f -rdfg query

9. Mask the newly created R2s to the DR ESXi hosts and share the devices information to the ESX team.

10. Perform a DR test again.

VMAX

vmax- R2 size not same as R1

Was this post helpful?