Link to home
Start Free TrialLog in
Avatar of stevenvel
stevenvelFlag for Australia

asked on

Write cache and performance hit.

ML 110G4 Server has built in RAID controller and RAID 1 is configured using 2x Seagate ST3160812AS SATA drives.
Recently I had issues with a bad spot on the RAID which I assume was the cause of (so far) two BSOD incidents.
I have run a RAID verify and it picked up and fixed 107 errors.
In the process I discovered I had been running the server with write cache enabled on the RAID controller but BIOS warning on POST recommends turning off write cache because the controller apparently can't handle this without the risk of data integrity issues.
The server has been running with cache write enabled for a long time (two years or more).
After disabling the write cache, server performance has dropped significantly to the point that it is almost unusable/unproductive (takes ages to boot up and CPU goes crazy).
The other thing to note is that for a long time (two years or more), one of the Seagate drives shows up on POST as 3.0GB/s and the other as 1.5GB/s.
I have not been able to work out why and if this is having a major impact on performance?
Please advice on possible fix.
How can I improve performance?
Should I just leave the write cache enabled and ignore the warning message on POST?
Should I try to get both drives to function at 3GB/s and if so how?
ASKER CERTIFIED SOLUTION
Avatar of charlestasse
charlestasse
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I would highly recommend that you invest in the best controller you can pay for, that has battery backup for this server
Avatar of stevenvel

ASKER

Yes you are right, the controller does not have a battery backup.
The server is due for upgrade but  I need to buy time.
In the short term, should I ignore the error and just enable the write cache?
I have a UPS always protecting the server.
Both drives are the same model (but not sure about firmware), how can I make them both run at 3GB/s?
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Even with UPS there is always the possibility for the server to shutdown unexpectedly, many things can cause this including power loss. Its up to you to determine if this is acceptable.
Server shutdown, such as caused by a system crash, will not result in loss of data when using a disk cache unless there is a loss of system power to the controller.  This could happen if the system power supply fails, for example, or if someone powers off the system manually.  

You should evaluate to what extent you can risk a data error.  Using a UPS and RAID significantly reduces the likelihood of unrecoverable data loss, as do regular backups.  
eager,
I have re-enabled write cache and I think performance is back to normal.
I checked the drives and confirmed that no jumpers are installed.
No jumpers should set the drives to work at 3GB/s but for some reason one drive displays as 1.5GB/s on POST.
It would be nice if they both show up and operate at 3GB/s for best performance but I don't know how to achieve this. Even if I could I am not sure if this would impact the RAID 1 setup (stuff it up)?
Not sure if it's possible to upgrade firware in the drives?
I may just have to leave this in the too hard basket and work on replacing this server.
This sounds like the best plan for you at this point Stevenvel.

Cheers
You might run smartctl on the drives. I don't know whether the transfer speed setting is one of the available settings for your drives, but the utility should tell you.  

Seagate has firmware updates for some of their drives.  Search on their website for your model.  

Different drive speeds will not affect whether the RAID works, but might affect the performance.  

Thank you both for your comments.