March 6, 2008 at 6:42 am
3+ hours for a 40GB backup? Is your NIC 100mb or 1GB? Regardless, that is out of control ... sounds like you may be experiencing some blocking. I assume you're using activity monitor to look at this SPID correct? Scroll to the right and check out blocked/blocking.
Also, what is the status of this spid? Running/Suspended/Sleeping?
March 6, 2008 at 6:51 am
Is it me or did my post end up above yours?
March 6, 2008 at 6:58 am
Its not you, Its definately above mine
Ive had a look , theres no blocking but the spid is suspended with a wait type of ASYNC_IO_COMPLETION
on the top row,BACKUPBUFFER in the middle row,IO_COMPLETION goes on and off on the last row where there is CPU activity
March 6, 2008 at 7:04 am
The_sponge (3/6/2008)
Its not you, Its definately above mineIve had a look , theres no blocking but the spid is suspended with a wait type of ASYNC_IO_COMPLETION
on the top row,BACKUPBUFFER in the middle row,IO_COMPLETION goes on and off on the last row where there is CPU activity
Hmm, these are types of problems I haven't had the privilege of experiencing ... usually suspended tasks are a direct result of an IO wait and your wait type definitely helps back that.
Have you attempted to back up the database locally or to another location to see what happens? My gut feeling is that you either have a network issue to your destination, or something up with the destination drive itself. I'm not much of a hardware guy, can't really shed more light than that at this time.
I'd say stop the process, try to back up locally and if successful, hopefully you at least can eliminate SQL from the equation.
March 6, 2008 at 9:08 am
Hi,
I may be missing some of the post, but from what I have been able to read, sounds like this may be disk subsystem related. What is your backup destination? What is the RAID configuation, if any? How many filegroups does the database(s) contain? What other activity have you observed on server?
Thanks,
Phillip Cox
MCITP - DBAdmin
March 6, 2008 at 9:44 am
This is in the error log:
Unknown,The operating system returned error 1450(Insufficient system resources exist to complete the requested service.) to SQL Server during a write at offset 0x000021fd852200 in file with handle 0x000019D8. This is usually a temporary condition and the SQL Server will keep retrying the operation. If the condition persists then immediate action must be taken to correct
Ive cancelled the backup and am going to try a new approach.
Im backing up to a remote path \\lon-nas01\fullbackups\backup.bak which is on a SAN storage on a RAID 1 storage. There are quite a few processes running so I think thats not helping things.
2 Filegroups, Primary and INdexes.
At this point the NAS is the only box on the SAN. I am going to connect the passive node of this DB that im trying to backup to the SAN
, do a failover,present a disk to the active node and do a backup.
March 6, 2008 at 9:55 am
As has been stated, you've added the network component into the mix. Backing up to a SAN is going to be slower than a local backup. Did you look at third-party tools (LightSpeed, SqlSafe, etc.)? These tool can compress the backup and speed up the process. Or can you backup local, maybe zip the file, and move the backup file to the SAN? How are you determining 40Gb short of completion?? If you're looking at the file on the SAN, my experience has been it'll read zero file size until completion occurs, so I wouldn't use that as a benchmark.
--FYI - the times of posting appear to be off as I posted after I read your post, which was time stamped at 10:27:12AM. Mine appears to have posted earlier. Something wierd going on here today!
-- You can't be late until you show up.
March 6, 2008 at 10:10 am
We use redgate on most of our servers. For whatever reason this server did not have it installed
I instaled it yesterday and tried to do a backup but it gave me connectivity problems.
The full backup size (.BAk) is 180gb. It was sitting at 140GB. But now I know not to trust SAN appearance
I dont have space locally for that size backup. I know redgate can compress by up to 5 times but I still dont have 30GB spare to do that, so I have no choice but to backup to the network.
March 6, 2008 at 10:18 am
I have to agree with Philip about the disk subsystem. Are any errors being raised? Can you backup a smaller database using Redgate? I don't believe it's the backup software. It's got to be hardware related or network. Get one of the infrastructure guys to help you trouble shoot (unless of course you wear that hat too!).
-- You can't be late until you show up.
March 6, 2008 at 10:27 am
Hi
Yesterday we started using a SAN and I am trying to backup a DB to it. The backup used to take around 2 hours to run but this backup has been running for over 3.
I have 1 SPID associated with the backup showing 3 rows in SP_WHO2
cputime Diskio
470204
873476349
10217190
The last row CPUTIME is increasing but all the others are still
Im trying to understand what is going on on a granular level here. I cant find any good documentation on how SQL does backups. There only seems documentation on "HOW TO" do backups.
The current state of the backup is that it is 40GB short of completion but its been running for 3 hours.The spid CPU is doing something -- but what???
Please can you point me to a good source of high level SQL technical documentation on backups if you know of any
Also how do people feel about SAN's. This is the first time im using one.
cheers
Chris
Viewing 10 posts - 1 through 9 (of 9 total)
You must be logged in to reply to this topic. Login to reply