mcanally@swift.cs.tcd.ie (07/13/90)
We have a DN10000 (wilde) and a 3550 connected to a campus wide ethernet. We have been unable to get OMNIBACK to restore any files from the exabyte tape drive which sits on the 10000. Omniback was installed sucessfully (v.1.1.p on the 10000 and v.1.1 on the 3550). The command given on the 10000 /etc/nbsrestore -input /dev/rmts8 -vol //wilde -trees / -into //wilde/sysprg/restore gave the following error: Op: CP0 Name: "" Command: "/etc/omniback/bin/dmux -dmux 4b8649d7f000.0d.00.02.54.dd.00.00.00://wilde -rsm 4b864973a000.0d.00.02.54.dd.00.00.00;//wilde -input /dev/rmts8" ?(spm) Local locksmith prevents locksmith login [error] (Fatal) Can't invoke '/etc/omniback/bin/dmux' on //wilde " local locksmiths only are allowed (process manager /remote process manager) Check if omniback is properly installed Has anyone come across this before, or has anyone got any idea how to fix this problem? We have had no backups running since we got the machine!!! Thanks in advance, , Maire Mc Anally e-mail: mcanally@cs.tcd.ie Computer Science Department, Trinity College, Dublin 2, Ireland.
thompson%pan@UMIX.CC.UMICH.EDU (John Thompson) (07/17/90)
> We have a DN10000 (wilde) and a 3550 connected to a campus wide ethernet. > > We have been unable to get OMNIBACK to restore any files from the exabyte > tape drive which sits on the 10000. Omniback was installed sucessfully > (v.1.1.p on the 10000 and v.1.1 on the 3550). 1) Get version 1.2 and 1.2.p. Please. I beg you. > The command given on the 10000 > > /etc/nbsrestore -input /dev/rmts8 -vol //wilde > -trees / -into //wilde/sysprg/restore > > gave the following error: > > Op: CP0 Name: "" Command: "/etc/omniback/bin/dmux -dmux > 4b8649d7f000.0d.00.02.54.dd.00.00.00://wilde -rsm > 4b864973a000.0d.00.02.54.dd.00.00.00;//wilde -input /dev/rmts8" > ?(spm) Local locksmith prevents locksmith login > [error] (Fatal) Can't invoke '/etc/omniback/bin/dmux' on //wilde " > local locksmiths only are allowed (process manager /remote process manager) > Check if omniback is properly installed Check whether Omniback is installed correctly. Is nbsrestore setuid root? If not, is the dmux executable by world? Are you using lprotect? Because you can run the front-end from just about any node, the dmux (and the mux, when you do a backup) is started up via a crp connection (as are the dw and dr programs). Specifically, you have the following programs running -- Backup Scenario: On front-end machine -- nbsbackup nbsbackup_netio On tape-drive machine -- nbsbackup_mux nbsbackup_mux_netio On machines being backed up -- nbsbackup_dw nbsbackup_dw_netio Restore Scenario: On front-end machine -- nbsrestore nbsrestore_netio On tape-drive machine -- nbsrestore_dmux nbsrestore_dmux_netio On machines being restored -- nbsrestore_dr nbsrestore_dr_netio Note two things: (1) If you're running the backup from the node w/ the tape drive, all 4 processes will be on that node; and (2) If you specified a dr_host / dw_host, the dr or dw processes will be on that host, regardless of what node's volume is backed up / restored (for each volume, of course). The '_netio' process appears to be forked by the "main" process, and handles the interprocess communication (????) between dr <--> mux <--> nbsbackup. The long and short of this (much longer than shorter) is that you must be able to CRP on to many nodes, _including_ _the_ _one_ _you're_ _on_. > Has anyone come across this before, or has anyone got any idea how to fix this > problem? We have had no backups running since we got the machine!!! Not specifically, but many general problems in the same vein. John Thompson (jt) Honeywell, SSEC Plymouth, MN 55441 thompson@pan.ssec.honeywell.com As ever, my opinions do not necessarily agree with Honeywell's or reality's. (Honeywell's do not necessarily agree with mine or reality's, either)