Horrible throughput after array rebuild

[lhosbrough]

Proliant ML 370 G4 with Ultrium 448 tape drive.

Had a drive fail in 3 drive array. Replaced it, and now have terrible throughput on backups. Tried doing Backup to disk, and throughput still terrible.

Smart Array 641: Array Diagnostic Report shows under--Storage Enclosure-Sense Bus Parameters

Connection Flag Running in LVD mode-Status informational
Using Id numbering-Status informational
U160 disabled-Status CRITICAL

Is this our problem? HP couldn't tell me what it meant. The array is made up of Controller Smart Array 641 Controller
Bus Interface 64-Bit PCI
Controller Location Slot 1

All 3 drives are as follows
Location Port 1 : SCSI ID #, Internal
Size 72.8 GB
Type Parallel SCSI
Actual Width Wide
Negotiated Width Wide
Transfer Speed ULTRA 320
Firmware Version HPBC
Model COMPAQ BD07287B4C
Status OK

Any assistance is greatly appreciated.

 

[spi200]

Hi


Maybe the disk is still rebuilding itself and is not yet back to a normal status. You can check in hte Array Configuration utility or insight Manager on the status of your array.
This error concern me as well "U160 disabled-Status CRITICAL" - not seen that one before but does not sound good and needs further invstigation.
Sounds like its running at speeds much lower than U320's are cabable of so probably some underlying problem. One of the other guys here may have seen this error before.

Regards

Dave

 

[TechCoordinator]

Which RAID are you running?
And, are all the drives the same speed?
And, what is the 'hardware' name of the replaced drive and the new drive?

 

[lhosbrough]

RAID 5.0
Drives all same speed (slight variation in model #), but all straight from HP, all FW upgraded on drives and tape drive.

Job summary is confusing. Different throughput rates on separate areas of backup job.

Set Detail Information - Backup \\MAILSERV\C:
Set type : BackupSet status : CompletedSet description : Backup 0006
Resource name : \\MAILSERV\C:Logon account : System Logon Account
Agent used : YesAdvanced Open File Option used : NoImage used : No



Start time : Wednesday, April 09, 2008 7:18:02 AMEnd time : Wednesday, April 09, 2008 1:26:24 PM
Files : 597,093Directories : 5,693Skipped files : 0Corrupt files : 0Files in use : 0
Byte count : 53,677,680,615 bytesRate : 138.00 MB/Min


Set Detail Information - Backup \\MAILSERV\Microsoft SQL Server "BLACKBERRY"
Set type : BackupSet status : CompletedSet description : Backup 0006
Resource name : \\MAILSERV\Microsoft SQL Server "BLACKBERRY"Logon account : System Logon Account
Agent used : YesAdvanced Open File Option used : NoImage used : No



Start time : Wednesday, April 09, 2008 1:26:25 PMEnd time : Wednesday, April 09, 2008 1:27:07 PM
Files : 0Directories : 0Skipped files : 0Corrupt files : 0Files in use : 0
Byte count : 28,003,031 bytesRate : 38.00 MB/Min


Set Detail Information - Backup \\MAILSERV\Microsoft Information Store\First Storage Group
Set type : BackupSet status : CompletedSet description : Backup 0006
Resource name : \\MAILSERV\Microsoft Information Store\First Storage GroupLogon account : System Logon Account
Agent used : YesAdvanced Open File Option used : NoImage used : No



Start time : Wednesday, April 09, 2008 1:27:08 PMEnd time : Wednesday, April 09, 2008 2:56:52 PM
Files : 0Directories : 0Skipped files : 0Corrupt files : 0Files in use : 0
Byte count : 27,474,371,864 bytesRate : 291.00 MB/Min

Backup of root C:\ is 138 MB, while Exchange is 260! Need to check what is being backed up in the root that is taking so long, I guess.

 

[spi200]

Hi

I have seen this sort of variations many times and I normally don't woory about it unless there has been an increase in the time overall backup job. Essentially some files come off the disk faster than others dependign on file size and amount of fragmentation of the volume etc - when the disk output can not keep up with the writing to tape then the tape drives stops "streaming" the data to tape and must do start\stop backup operations.
Some of your thru puts numbers look fine to me, try just backing up some simple folder\files from a file server, not a database of any kind.
I also assume that no other changes were made to the backup software or controler card for the tape drive etc as even a chnage in scsi id's can have an effect on performance.


Regards

Dave

 

[burtsbees]

What part number are all 3 drives? Did you replace the drive with an Ultra 3 or 320, 10K or 15K? I know what you posted, but what do you mean slight variation in model type? Did the array finish rebuilding before you tried a backup?

Burt

 

[lhosbrough]

All drives are same speed and specs (320) Array had finished building before any backups were attempted. The backup now takes twice as long for half as much data.

 

[technome]

What is the tape drive attached to ?


........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[xmsre]

Are all three drives using identical firmware?