Amanda Group, I've been running Debian Testing for almost 4 months now without any issues with chg-scsi binary and my quantum superloader 16 tape library LTO-3.
Couple of days ago chg-scsi has been unable to access the medium changer. Details of system and software: Kernel 2.6.17.7 with SCSI generic support (worked with this kernel) Debian Linux 3.1 Testing Amanda-server 2.5.1p1-2 Amanda-client 2.5.1p1-2 Amanda-common 2.5.1p1-2 Checking /proc/scsi/sg/device_strs CERTANCE ULTRIUM 3 1856 QUANTUM UHDL 0031 dmesg output <omitted> ch 1:0:5:1: Attached scsi changer ch1 ch 1:0:5:1: Attached scsi generic sg7 type 8 ls -al /dev/sg7: crw-rw---- 1 backup root 21, 7 2005-02-25 22:38 /dev/sg7 mtx -f /dev/sg7 status (works correctly) Storage Changer /dev/sg7:1 Drives, 16 slots ( 0 Import/Export) Data Transfer Element 0: Full (Storage Element 6 Loaded): VolumeTag = MBW.. Storage Element 1: Full :VolumeTag=MBW... ... ... ... Storage Element 16: Full :VolutmeTag=MBW... also mtx -f /dev/sg7 next (works) Unfortunately: Output of chg-scsi -info as follows: chg-scsi: open: /dev/sg7: No such file or directory <none> open: /dev/sg7: No such file or directory I've tried running both as root and the backup user. I've changed user/group permissions. I have another changer device an Exabyte 7 tape changer library at /dev/sg4 and if I change the amanda.conf/changer.conf file to point to SG4 chg-scsi -info reports: 13 16 1 1 This is incorrect, because the changer arm is not at the 13th slot, there isn't a maximum of 16 tape library, tape is not loaded into the drive, and there is no barcode reader on SG4. Any other /dev/sg[0-9] other than 4 and CHG-SCSI seg faults. I have cycled the server, tape drive/changer(s) with no success, changed Kernel to 2.6.18.2 with no success. Has there been any changes to chg-scsi from 2.5.0 to 2.5.1? The only thing I can think of is that I did a package upgrade and that's when things went downhill. Unfortunately none of my logs are reporting software changes. :( Also, no noteable changes in the bootup messages have been found to indicate tape changer behavior. Anyone have a similar issue or am I to now change to chg-zd-mtx and have to recreate the label file of the barcode database (such a drag!) Thanks for any assistance.