I am new to solaris and I replaced a faulty tape drive sun DLT7000
But, I am getting the follwoing error when system reboots
ltid deamon error drive index 1 is not correct, drive name /dev/rmt/2cbn is incorrect no such file or directory.
I have two drives the other one is /dev/rmt/0cbn, But backup cannot done due to this error.
what has gone rong and how to solve?
do a
devfsadm -Cv
and after this a
devfsadm -v
or a
reboot -- -r
the new drive should have a new devname and 0 and 2 where the names before the replacement...
when you attached the new tape drive, all device paths that the os is "using" for the new tape drives are different then what you previously had. 0cbn is no longer available. therefore, either do what duke has said to clean up the device tree or figure out what the tape mount is using. its been awhile since i've dealt with the wonderful world of backup administration but if you are using something like legato networker, you can use inquire to list what scsi devices are attached and seen by the os. you can find your tape device path via that.
Yes, after running the command reboot -- -r I could see now the creation of new devpath /dev/rmt/1cbn, and through the netbackup software i reconnected the missing path to this one. Things are ok so far and the backup started as expected.
I realy appreciate your help.
---------- Post updated at 01:00 AM ---------- Previous update was at 12:58 AM ----------
I have another request collegues.
I am interested to replace a faulty hard disk which is not mirorred. it is containing ex ./dev/vx/rdsk/Vol1 file system.
But i have some ideas in my mind and I wish if you can correct me if wrong
1- backup the files system using ufsdump utility to a tape
ufsdump 0uf /dev/rmt/0m /Vol1
2- umount the file system
3- Replace the faulty hard disk.
4- create new file system on the new hard disk
newfs /dev/rdsk/c1t1d0s0
5- Mount /Vol1 on the ne file system.
mount /dev/dsk/c1t1d0s0 /Vol1
6- Restore from backup.
ufsrestore -rvf /dev/rmt/0m
Are these steps correct particularly point no 4.?
Are you sure what you are doing? Your filesystem /vol 1 is mirrored under veritas volume manager so you need not do these steps. You can replace the faulty disk online by using the vxdiskadm
Yes I am sure . Because when this disk went faulty the Os couldnt see the mentioned file system and if it was mirorred this would not happen. More over the fsck command couldnt stat the mentioned file system.
what I did is : I physically pull out the disk and inserted back after 60 seconds, rebooted the system and OS could see the file system again and was able to run fsck to repair the file system.
there should be no need to reboot the os for the disk to be seen by solaris. you should be able to do this via vxdiskadm.
however, if it were unmanaged by veritas, luxadm or cfgadm should do the trick to replace the failed drive.
Thanks pupp, But as I said before I am new in solaris and was not having the experience of vxdiskadm. Now you are mentioning very intersting tricks ( luxadm & cfgadm ). How to use them in mycase? and what are the steps.?
Please advice.
Look at the man pages for cfgadm and luxadm. These commands only allows the OS to see the disk. you can remove and insert disks using this command. But to sync the disks, you still have to use the vxdiskadm menu.
#vxdiskadm ->choose appropriate option to replace failed disk