Hello again. At the risk of being banned from the list forever for unending questions, here we go:
I've narrowed my problem to dump. I tried running dump manually on the device (/dev/sda1) that refuses to backup and it is apparently getting very upset. It will read a seemingly arbitrary amount of data from the filesystem, then just block forever. However, the amount it reads is sometimes consistent, so there's something magical about the amounts. Sometimes it will read to the end of the partition and then stop, at which point, one of the child dump processes goes zombie, or it will read a gig, or maybe 600 mb, or maybe a little over 100 mb, etc. Consecutive runs do not always yield the same results. On a suggestion from another sysadmin, I specified the tape density and size. Upon doing that (specified each at 65534), I got it to run all the way to the end of the filesystem, but then, once again, it just stops there forever. This is apparently why my amanda backups were failing, dump timed out and so amanda killed the dump processes after a half hour or so. So, do you guys have any idea what might be causing dump to behave in this manner? Oh, as a side note, I was originally using dump version 0.4b21-3 and have upgraded to dump version 0.4b24-1, with the same results on Redhat 7.1. Oh, the filesystems that I'm trying to backup are coming off of a RAID array. I don't know that that will make any difference, but perhaps it will. -Tired in Philadelphia
