Unsolved

This post is more than 5 years old

165 Posts

2249

July 23rd, 2008 10:00

New Devs in a SRDF device Group

Hi

I have a SRDF device Group configured (tasedocuenta ), I'am Using Storage Foundation (Veritas Volumen Manager) to manage the volumes, I just added 5 DMX-4 disks to Veritas Volumen Manager Control.

Before I added the disk I was replicating from one Data Center to other data Center By using SDRF.

the question is:

I want to add the new devs to my existent srdf device group (tasedocuenta ) in order to have all my devices replicating from Mexico City to Queretaro City.

My script is as follows

symld -sid 3875 -g tasedocuenta add dev 0666
symld -sid 3875 -g tasedocuenta add dev 0667
symld -sid 3875 -g tasedocuenta add dev 0668
symld -sid 3875 -g tasedocuenta add dev 0669
symld -sid 3875 -g tasedocuenta add dev 066A

File devtascb

0666 0300
0667 0301
0668 0302
0669 0303
066A 0304

symrdf createpair -f devtascb -sid 3875 -rdfg 1 -type rdf1 -invalidate R2
or
symrdf createpair -g tasedocuenta -f devtascb -sid 3875 -rdfg 1 -type rdf1 -invalidate R2


could you tell me if is there something wrong in the script?

best regard

124 Posts

July 23rd, 2008 23:00

Hi Guanabacoa :)
Ok, seems like you running on dynamic SRDF !!
both the scripts will work.. but they would work differently

script :symrdf createpair -f devtascb -sid 3875 -rdfg 1 -type rdf1 -invalidate R2 - Will create the rdf devices but will not add them to device group nor will start the device sync.
you will have to manually add them by using :
symld -sid 3875 -g tasedocuenta add dev 0666

Now the below script :
symrdf createpair -g tasedocuenta -f devtascb -sid 3875 -rdfg 1 -type rdf1 -invalidate R2
will create the rdf devices R1 & R2 and add them to your existing device group tasedocuenta. thats fine...

But inorder to start the replication you need to run the establish command on the devicve file.

symrdf -file -sid --rdfg 1 establish ---> this will start the syncing of the devices. and then you can always check the srdf sync status by querying the group.

Let me know if this was of any help :) Cheers !!

7 Posts

July 24th, 2008 09:00

Taking this a step further, I want to associate the R2 with a R1/BCV. From what I've read I can do this with the command: symbcv -g DG -rdf associate dev XXXX and then use the command: symmir -g DG -rdf establish -full to create my initial copy.

I need some guidance on how to setup the 2nd hop. How would I associate and sync the R1/BCV with the R2 on a third frame? How does the R1/BCV know which partner R2 device to use? I will be using SRDF/A for the 2nd hop since the target frame is located across the country.

6 Operator

 • 

2.8K Posts

July 24th, 2008 09:00

While a TimeFinder relation is dynamic (you can establish any STD to any BCV, size permitting) an SRDF relation is a lot less dynamic then what you may think :D

If you have an R1/BCV device you already have its R2 device and don't need to associate them .. Maybe I misunderstood your question, however it looks like you want to associate BCV devices at R2 side of the link and manage them while staying at the R1 side of the RDF link. May I suggest you to open a new thread and explain better/in a different way what do you want to achive ??

This will allow you to award points to the best answer thus allowing others to find easily the answer to your (and maybe their) problems. :D

Message was edited by:
Stefano Del Corno

124 Posts

July 24th, 2008 21:00

Hi guanabacoa :)
please mark this answered, if it was !!

165 Posts

July 25th, 2008 16:00

Hi

I got trouble

I tried to run this command but I got this error


symrdf createpair -f /home/jecortes/devedocta/devtascb_2 -sid 3875 -rdfg 1 -type rdf1 -invalidate R2

The devices (invalidate target) must be write disabled when creating RDF devices with the invalidate flag

I writed disable my targets

symdev ¿sid 0766 write_disable 0300 ¿sa 1D ¿p 1
symdev ¿sid 0766 write_disable 0301 ¿sa 1D ¿p 1
symdev ¿sid 0766 write_disable 0302 ¿sa 1D ¿p 1
symdev ¿sid 0766 write_disable 0303 ¿sa 1D ¿p 1
symdev ¿sid 0766 write_disable 0304 ¿sa 1D ¿p 1

but i got this output:

0300 Not Visible ***:* 01A:D3 RAID-5 N/Grp'd RW 8632
0301 Not Visible ***:* 01A:C4 RAID-5 N/Grp'd RW 8632
0302 Not Visible ***:* 01A:D3 RAID-5 N/Grp'd RW 8632
0303 Not Visible ***:* 01A:C4 RAID-5 N/Grp'd RW 8632
0304 Not Visible ***:* 01A:D3 RAID-5 N/Grp'd RW 8632

I can't see the devices in WD mode

Does any one know what is my mistake?
is there another way to do that?


Thanks in advance

11 Legend

 • 

20.4K Posts

 • 

87.4K Points

July 26th, 2008 03:00

what if you try to disable without specifying FA:

symdev ¿sid 0766 write_disable 0300

6 Operator

 • 

2.8K Posts

July 26th, 2008 06:00

As Mr Dynamox said, you write disabled your devices only on a single path.. Try with a

symdev -sid xxx show 0BAD

and look at bad device to find if it's really write disabled on desired port.

6 Operator

 • 

5.7K Posts

August 1st, 2008 02:00

There are 2 ways for this:
1) make sure you do the createpair when the targets are unmapped
2) write disable the targets on all FA's they're mapped to

Well, a 3rd was is to unmap them before doing the createpair, but that's a time consuming action.

2 Intern

 • 

1.3K Posts

August 3rd, 2008 13:00

both first and third option says unmapping.. Did u mean unmasking in one of the cases?

6 Operator

 • 

5.7K Posts

August 12th, 2008 14:00

No, simply write_disable them OR unmap them. You could unmask them first, then unmap and after the create_pair you need to map and mask again. I'd say: skip the unmask and mask commands, since after all this you;ll need to have them masked again anyway.
But if you want to have the vcmdb consistent with anything that's mapped, you'd better unmask and mask again as well ;)

I would simply write_disable them, that's waaaaaaaaaaaaay quicker.

6 Operator

 • 

2.8K Posts

August 26th, 2008 03:00

Did you solve your troubles ??

6 Operator

 • 

5.7K Posts

August 26th, 2008 03:00

I can see my 3 options are confusing.... Now, 3 weeks later I'm reading my answer and I don't understand it anymore myself ;)

I guess you were right: 2 options: write_disable or unmap.

165 Posts

August 27th, 2008 07:00

Thanks Everything is OK.

6 Operator

 • 

5.7K Posts

August 27th, 2008 07:00

Please mark the question as answered and reward any correct or helpful answers please ;)
No Events found!

Top