On several occasions I've had the usual CRC errors, and have been able to remedy them with new tapes, cleaning the tape heads etc etc. Now however, I have a very strange re-occurring CRC error.
I am running BE 8.5 on Windows NT with a robotic library with 2 tape drives in. On this particular backup server I am running 4 jobs, each one backing up a different server. Now, 3 out of the 4 servers backup with no problems. However, the 4th keeps getting a CRC error. I swapped for brand new tapes, and cleaned the drive and tried moving the job so it backs up on the other drive in the library. Each time, a CRC error. This would lead me to suspect that there was a corrupt file on the server in question. However, the CRC error occurs at different points in the backup each time.
Since the server (linux server by the way) has 4 logical disk partitions, I tried creating 4 separate jobs, 1 for each disk, hoping to narrow down the location of the corrupt file(s). This time however, everything ran with no problems. As soon as I combined the jobs back again, the CRC error came back.
Does anyone have any suggestions as to why this should keep happening? I could leave the backup of this server as 4 separate jobs, but it gets quite untidy and uses more tapes than are really necessary. I am aware that changing the block and buffer size of the tape drives is a potential solution, but I’d rather not as this is going to cause no end of problems when I come to restore data from tapes written with a different block size on different backup servers. Quite likely there are no other solutions to this problem, but does anyone know why its actually happening? If I knew the cause I might be able to work around it somehow.
Many thanks in advance,
Iain
I am running BE 8.5 on Windows NT with a robotic library with 2 tape drives in. On this particular backup server I am running 4 jobs, each one backing up a different server. Now, 3 out of the 4 servers backup with no problems. However, the 4th keeps getting a CRC error. I swapped for brand new tapes, and cleaned the drive and tried moving the job so it backs up on the other drive in the library. Each time, a CRC error. This would lead me to suspect that there was a corrupt file on the server in question. However, the CRC error occurs at different points in the backup each time.
Since the server (linux server by the way) has 4 logical disk partitions, I tried creating 4 separate jobs, 1 for each disk, hoping to narrow down the location of the corrupt file(s). This time however, everything ran with no problems. As soon as I combined the jobs back again, the CRC error came back.
Does anyone have any suggestions as to why this should keep happening? I could leave the backup of this server as 4 separate jobs, but it gets quite untidy and uses more tapes than are really necessary. I am aware that changing the block and buffer size of the tape drives is a potential solution, but I’d rather not as this is going to cause no end of problems when I come to restore data from tapes written with a different block size on different backup servers. Quite likely there are no other solutions to this problem, but does anyone know why its actually happening? If I knew the cause I might be able to work around it somehow.
Many thanks in advance,
Iain
