This post is more than 5 years old
13 Posts
0
1364
October 5th, 2007 04:00
Expansion of logdrive on SQL cluster
Hi we are running a windows 2003 server cluster, with SQL server 2005 enterprise ontop of it, in an Active/passive configuration
Im running into problems with out logdrive is too small (50Gigs) and i want to extend/expand that space.
What is the correct way to do this?
Regards
Christian
Im running into problems with out logdrive is too small (50Gigs) and i want to extend/expand that space.
What is the correct way to do this?
Regards
Christian
No Events found!
AranH1
2.2K Posts
0
October 5th, 2007 11:00
On powerlink go to:
Support > Technical Documentation and Advisories > White Papers > Configuration/Administration
and look for the CLARiiON Best Practices for Fibre Channel Storage document. It describes the file system alignment issue with RAID stripes and how to correct.
For your post though this is not an issue since you most likely did not create the partitions on your LUNs with a correct file system alignment offset to match the RAID stripe.
The migration process does not require any extra steps on the array if you are migrating to a larger LUN. As mpi and dynamox pointed out you will just need to extend the partition to take advantage of that space afterwards.
Aran
mpi2
2 Intern
•
215 Posts
1
October 5th, 2007 05:00
Do you have clones or snaps running on it (e.g. RM)?
You have two options:
-) buy disks, create a Lun on it and migrate the log lun to it.
-) buy disks, and expand your Lun with a striped expansion.
After the migration/ expansion. You have to rescan the drives. and then expand the fs with diskpart on the active node. after that test the failover. if you have any storage software using the lun (clone, snap, mirror) you have to stop this before you can expand the disk.
best regards
Manfred
Darkingdk78
13 Posts
0
October 5th, 2007 08:00
will the migrate to the raid1/0 be seemless for the server?
If i migrate to a larger volume that the original, will i have to do it in more than one step?
/Christian
Darkingdk78
13 Posts
0
October 5th, 2007 09:00
dynamox
11 Legend
•
20.4K Posts
•
87.4K Points
0
October 5th, 2007 09:00
dynamox
11 Legend
•
20.4K Posts
•
87.4K Points
1
October 5th, 2007 11:00
i was going to say create alignment within Navisphere on the target lun ..but then i saw this article emc131645 ..that scared the **** out of me.
dynamox
11 Legend
•
20.4K Posts
•
87.4K Points
0
October 5th, 2007 12:00
AranH1
2.2K Posts
0
October 5th, 2007 12:00
But if I am reading that right it is with an offset created on the array not when the file system was created on the host.This could be why in the EMC documents they recommend using the process of creating the partition on the host to set the alignment offset instead of the LUN offset on the array.
Creating the alignment offset on the host during the partition creation process would also simplify any future migrations, clones, etc.. of the LUN right?
Darkingdk78
13 Posts
0
October 5th, 2007 13:00
Thanks for all the help, i will give it a go after a good nice backup of all our drives
Regards
Christian
AranH1
2.2K Posts
0
October 5th, 2007 14:00
You are correct, the issue was with an offset configured on the LUN within the array. I have never had any issues migrating LUNs either without an offset or with an offset configured during the partition creation on the host.
One more thing though, the migration can have a performance impact on a host and since this is a sql log LUN I recommend you kick this off during a low i/o period.
Aran
Darkingdk78
13 Posts
0
October 8th, 2007 04:00
No really it worked super, thank you all for the great help, now our sql logs are running on a... by all means faster RAID group than before.
Raid 1/0 is certainly faster for logs than Raid 5.
Hugs all around!
Regards
Christian