JerryK440
asked on
Disk Corruption - Event ID 55 and KB 932578?
We have a Windows Server running 2003 with SP2 applied. Our environment is completely Terminal Server oriented. (ie. All of our staff connects to the Terminal Server to do their work, nothing saved on locals.)
Periodically we are getting Event 55's on one of the two physical disks. This has been happening for about 16 months now. Originally it was on a "real" server. Then we imaged it and set it up on a VMware server. Still had the issue. We purchased a new server and completely reinstalled the OS and applications. Only thing copied from the old system was the data. The problem has persisted.
I have been researching this issue and the only thing that comes close to what we have going, seems to be addressed in MS KB 932578. http://support.microsoft.com/kb/932578
However, that is for systems with allocation units smaller then 4096. Unless Windows is reporting it wrong, our server has a size of 4096.
Everything else I have found, has been older then what we have already installed.
We don't do anything with double-byte characters either.
It always seems to happen around specific times/days of the month. Though nothing is consistant with the programs running at the time. All that seems consistant is that it is heavier disk io.
So.... my question is (finally)... Is it possible that the fix in KB 932578 could fix this issue, even though my unit size is 4096? And is there any downside from trying this?
Thanks,
Jerry
Periodically we are getting Event 55's on one of the two physical disks. This has been happening for about 16 months now. Originally it was on a "real" server. Then we imaged it and set it up on a VMware server. Still had the issue. We purchased a new server and completely reinstalled the OS and applications. Only thing copied from the old system was the data. The problem has persisted.
I have been researching this issue and the only thing that comes close to what we have going, seems to be addressed in MS KB 932578. http://support.microsoft.com/kb/932578
However, that is for systems with allocation units smaller then 4096. Unless Windows is reporting it wrong, our server has a size of 4096.
Everything else I have found, has been older then what we have already installed.
We don't do anything with double-byte characters either.
It always seems to happen around specific times/days of the month. Though nothing is consistant with the programs running at the time. All that seems consistant is that it is heavier disk io.
So.... my question is (finally)... Is it possible that the fix in KB 932578 could fix this issue, even though my unit size is 4096? And is there any downside from trying this?
Thanks,
Jerry
ASKER
Thanks for the suggestions, Shreedhar. I have already done that several times in the past 16 months.
In fact I am running yet another chkdsk /r as I type this, since ironically the issue happened again shortly after I posted my question.
Yesterday, we had switched over to our backup server when the issue cropped up. So the hardware is different, yet the issue stays the same. Though in this case it was corruption on drive C, rather then drive D yesterday.
CHKDSK /R usually fixes a couple items, then everything is fine for a time.
Like I said, the only thing that seems consistant is that the issue happens when there appears to be high disk io. Which seems to point me to the NTFS.SYS file. I am assuming that all disk IO would go through that driver, regardless of the physical drive being written to.
In fact I am running yet another chkdsk /r as I type this, since ironically the issue happened again shortly after I posted my question.
Yesterday, we had switched over to our backup server when the issue cropped up. So the hardware is different, yet the issue stays the same. Though in this case it was corruption on drive C, rather then drive D yesterday.
CHKDSK /R usually fixes a couple items, then everything is fine for a time.
Like I said, the only thing that seems consistant is that the issue happens when there appears to be high disk io. Which seems to point me to the NTFS.SYS file. I am assuming that all disk IO would go through that driver, regardless of the physical drive being written to.
Hi,
Update the RAID Contoller drivers.
---------
Shree
Update the RAID Contoller drivers.
---------
Shree
ASKER
Thank you again for the suggestions, Shreedhar. As I stated before though, have already done that.
Drivers/Firmware have all been updated for everything in the PC. (Even CD drive that is never used on the server.)
The problem has persisted, and has manifested itself on four separate sets of hardware. Always under "heavy" load. Heavy in quotes because the current server is not even working up a sweat. The utilization percent is rarely over 10% for most items. (Aside from one CPU core now and then when one user does "compiles" financial statements for a client.)
Which is why I am looking at MS KB 932578. http://support.microsoft.com/kb/932578 That patch seems to describe the issue precisely, but is for systems with a cluster size of smaller then 4096. Our server has a size of exactly 4096.
Would there be any adverse effects in installing this patch on a server with a cluster size of 4096?
Drivers/Firmware have all been updated for everything in the PC. (Even CD drive that is never used on the server.)
The problem has persisted, and has manifested itself on four separate sets of hardware. Always under "heavy" load. Heavy in quotes because the current server is not even working up a sweat. The utilization percent is rarely over 10% for most items. (Aside from one CPU core now and then when one user does "compiles" financial statements for a client.)
Which is why I am looking at MS KB 932578. http://support.microsoft.com/kb/932578 That patch seems to describe the issue precisely, but is for systems with a cluster size of smaller then 4096. Our server has a size of exactly 4096.
Would there be any adverse effects in installing this patch on a server with a cluster size of 4096?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Thanks Shreedhar. I will be trying that tonight.
Hopefully all will go well, and our users will finally have some peace to work in!
Hopefully all will go well, and our users will finally have some peace to work in!
Jerry, did this solve your issue? We are experiencing the same thing on our terminal servers!
First run chkdsk on all the drives and if it reports error then backup the entire data and system state of the server.
After that run chkdsk /f to fix the errors.
Check with your hardware vendor for any updates for the BIOS, RAID Controller If any and NIC. Also verify the server HDD for any issues.
I hope this helps,
Shree