This post is more than 5 years old
75 Posts
0
2834
What is this? Software raid on windows?
Hi guys,
I don't know much about windows, so I need your help to figure out if this setup is actually supported by EMC and supposed to work. The problem is that backups hangs after a long while when they try to backup a server containing their home directories. I've tried several things to solve this, but I'm beginning to suspect that the problem lies in the fact that they have their AFTD's on this device. If I do the same backup direclty to tape, it works fine.
They have 3 AFTD's on this, and it's only the backup of the fileserver with the homedirs that fails.
Some screenshots:
So, my questions are:
1. What is this, and what is it called?
2. Does it work?
3. Is it supported?
Thanks guys,
Mark_Bellows
240 Posts
0
April 13th, 2011 07:00
Tony,
I am glad that we were able to help you with this issue!
Mark
tonyalbers
75 Posts
0
April 6th, 2011 06:00
NW 7.5.3.2 on Win 2003 btw.
/tony
ble1
2 Intern
2 Intern
•
14.3K Posts
1
April 6th, 2011 09:00
If I understood correctly, your backup fails after they start their backup. Why two backups in first place?
tonyalbers
75 Posts
0
April 11th, 2011 02:00
The setup is made so that a backup of the fileserver ends up in "BACKUP_DISKCACHE" which contains some AFTD's. The data is then staged to tape after a while.
The backup starts and runs normally, but when it's about to finish, it suddenly stop with an error message stating:
71909:save: Failed to close save session: cannot find ssid 38fd5e65-00000006-9f8d6985-4d8d6985-8c661700-0e533ba4 to update
: Failed to close save session: cannot find ssid 38fd5e65-00000006-9f8d6985-4d8d6985-8c661700-0e533ba4 to update
The mentioned ssid exists nowhere in networker, so I guess it's the one its actually creating while running the backup.
If I start the exact same backup job, but send the data to tape, it finishes without errors.
So, there is only this one backup, we just tried to use tape instead to be able to exclude the client from the list of possible errors.
/tony
tonyalbers
75 Posts
1
April 11th, 2011 03:00
The backup is just around 2TB. A lot of small files(since it's homedirs).
Tape drives/library and disk system are on different controllers.
The 24TB volume consists of a number of LUN's from a HP EVA SAN which is mounted as D: and C:\BACKUP_DISKCACHE ( I asked them to mount it on D:, but I haven't used it yet) They have used C:\BACKUP_DISKCACHE all the time.
/tony
ble1
2 Intern
2 Intern
•
14.3K Posts
1
April 11th, 2011 03:00
How big is this backup? One thing which is different for sure is controller (as disk and tape should not share one). In such case, does your disk share same controller? I'm a bit confused with those snaps since I see 24TB disk reported as D: but then again it seems to be showed under C: drive.
Anyway, can you start your test in a way that you test it against something smaller so you can eventually repeat tests with some additional debug options? Also, does this happen no matter what backup level you use?
Mark_Bellows
240 Posts
1
April 12th, 2011 11:00
I think Hrvoje is on the right track. If I am reading this correctly, your diskcache volumes are 2TB in size. Although the data may be 2TBs in size, there is some overhead that goes with the data and it ends up with more than 2TB of total data written to the volume. I see that these volumes are in RAID set together, but I am not so sure the configuration is set correctly. I say this because one thing that I found interesting is that, looking at the disc managemnet picture, is that the DiskCache volume shows a NEGATIVE 82% free space. This tells me there is an issue with that configuration.
Please review this article to see if it helps you resolve the issue:
Adding support for more than eight LUNs in Windows Server 2003 and in Windows 2000
Also, if the backend disk unit (HP MPIO) can present the LUNs as one unit, that would remove the configuration from the server and place it on the storage unit.
I hope this helps!
Mark
tonyalbers
75 Posts
0
April 13th, 2011 04:00
Hi guys,
Thanks for the advice, I'll try to do a test with a smaller backup and try to have the HP MPIO create the volume for me.
I'll get back to you with the results.
/tony
tonyalbers
75 Posts
0
April 13th, 2011 06:00
Hrvoje and Mark, you helped me a lot on this one. I'm very thankful for your advice and suggestions.
I just got an email from the customer who wrote that they would like us to close the ticket because it seems that this errors was caused by a fault in their SAN. I guess this also caused the strange -80% free reporting in disk management that Mark pointed out.
Again: Thanks guys.
Mark: Please post a reply so that I can credit you for a correct answer.
/tony