CRC Errors / Soft Write Errors / Backups are not reliable

[Jon_Ziminsky]

Hello!

We are running BackupExec v10.1d(5629) on Windows 2003 connected to a Certance 8 slot LTO Autoloader. Everrything was running fine with few to no failures up until about December 16, 2006. At that point in time everything seems to have fallen on its face.

I am getting a myriad of symptoms, and am not having luck in narrowing the problem down any further than where i am now. Here is a list of things that i have tried to no avail:

1). New media. I have been using the same brand (TDK) of media for years with no problems. My first assumption of what was causing the CRC errors was the media itself, so i ordered in another batch. Still recieving errors.

2). Repair installation of Backup Exec. This did not help the problem.

3). Latest greatest firmware/drivers for our Certance Autoloader. This did not help.

3). New cleaning tape for tape drive. Ran it through cleaning cycles multiple times. This did not help.

4). Hotfix 45. I tried this because i was also having problems with the "Recovery" user showing up in my reports, and with jobs hanging after they had been cancelled.

5). I have ran tests on the library itself, and it is returning all good.

6). Delete and recreate all jobs, policies and selection lists. Recreate all jobs, policies and selection lists. This did not help.

I did notice on the media that i am experiencing a HIGH rate of soft write errors. I put three brand new pieces of media in last night and have the follwoing after the jobs tried to run... First piece has 152641 soft write errors, Second piece has 124973 soft write errors, and the third piece has 60986 soft write errors and 2 hard write errors.

With write errors that high i am tending to feel as if my tape drive is failing me.

Any thoughts? Thanks in advance.


Jon Ziminsky
CPS Distributors, Inc.
Denver, Colorado

[Hywel_Mallett]

It sounds similar to a problem I had with a HP tape drive - performance was poor. Investigation showed that it was poor because it was having to correct so many errors!
I logged it with HP, and they replaced the drive. The replacement has been fine (touch wood).
It does sound like a tape drive issue to me.

[Jon_Ziminsky]

Well, my theory of a bunk tape drive has been thwarted... I ran the diagnostic utilty from Certance 5 times with varying data amounts and was 100% succesful every time. The library is functioning properly, and the drive is able to compress data with no problems.

So now i am truly at a loss. The CRC errors and Soft Write errors had me convinced that my tape drive was shot.

The next step i am trying is repartitioning the drive. I am going to take it back to 5slots, 2slots and a single slot for cleaning, as opposed to the 2x(3slots) 2x(1slot) i currently have. I am doing this to try and rule out a lack of space on my media.

The media that i used last night was showing that each piece had at least 20 gigs free on it. One tape did have a "End Marker Unreadable" notation on it.

I am posting as much information regarding this as i can possibly think of to try and maybe match similarities that one of you may be facing currently, or may have faced in the past.


Thanks for taking the time to read this issue.




Jon Ziminsky
CPS Distributors, Inc.
Denver, Colorado

[Jon_Ziminsky]

Does Symantec actually "moderate" these forums???

If so, what is the average time to receive a response? 3 months? Personally i consider my backups as a crucial part of our organization. It is ashame you all do not consider customer support to be in the same realm.



Jon

[Hywel_Mallett]

Please update us on the issue.

;)

[Francois_Steyl]

Hi Jon.

CRC errors normally indicates to the hardware. As a test, can you take one of your backup Jobs and run that to disk?

Try to use the same job and allocate some disk space to backup exec to use as a device. If you are still getting all the soft errors then at least we know it is not the hardware device.

Do the same test then by using a different backup selection list, and see what the errors are.

Hope this helps.

[Hywel_Mallett]

In my experience soft read write errors will 99% of the time be caused by worn media (which you have already ruled out), or faulty hardware or a SCSI misconfiguration.
To give you some statistics for comparison, my best-performing tape has 0 soft write errors after 352 hours of use. My worst performing tape has 16 soft write errors after 401 hours of use.
I've only seen stats as bad as yours with a faulty drive.
I haven't used the Certance diagnostics. I wonder if they are any good?
Just in case it's something to do with the compression on the drive, have you tried running a job with no compression?

[Jon_Ziminsky]

This problem is still occuring. I had a few days of succesful backups, but last night they failed again becuase of "CRC Errors".

I have resigned myself to the fact that it is my tape drive... We are moving to a PowerVault LTO3 library. I hope this clears up the problems i have been having.


Jon

PS. It sure is good to know that Symantec looks at these forums a couple times a year...