steven@batserver.cs.uq.oz (Rm 303 Phone 3973) (09/28/89)
At the University of Queensland, we have 7 DN3500 Apollo Workstations (4 of which are disked) running sr10.1 in a bsd environment. We would like to back these up over the network to a remote host (which happens to be a Sun) with a 9 track tape drive. What we are looking for is a backup utility (other than wbak) which will be able to do full and incremental backups of our disks. Apollo's OmniBack product does not allow for the tape host to be a non-Apollo machine and thus is not usefull in this instance. The 'tar | rsh host dd' system does not handle multiple tapes and requires a breakdown of the file system. It also does not cope well with incremental backups since it cannot handle deletions. Does anyone know of, or use a system for backups (commercial or inhouse) that would work well in this type of environment. Thanks in advance. Steven __ Steven Widmayer Asst Systems Programmer University of Queensland. steven@batserver.uq.cs.oz Better to burn out than fade away.
abair@turbinia.oakhill.uucp (Alan Bair) (09/30/89)
I know you said you did not want to use wbak, but take a look at the -remote(?) option. With SR10.1, you can run wbak and send the output to the tape on the Sun. I have used it and it works, but I really want it for our DN10000, but its not there yet. Actually before this option was in place, I grabbed a PD verison of a set of rmt routines, the code that does remote tape ops, and made my own little cat to run 'wbak -stdout | rcat ...'. This worked fine to 1/2" tape. I need to get it setup to use our 8mm, then no more tape swapping, even for full backups. Hope you can use this. Actually I would rather use the Unix dump and restore programs, but Apollo does not support them. Alan Bair SPS CAD Motorola, Inc. UUCP cs.utexas.edu!oakhill!turbinia!abair
phcalamai@water.waterloo.edu (Paul H. Calamai) (09/10/90)
HELP!!!! There something that I'm attempting with backups that I'm still having problems with. Currently the backup does not include the registry information in /sys/registry. The reason for this is the registry has to be put into maintenance mode before it can be copied. This can only be done by the owner (root). I tried running the following script # Putting master registry in maintenance. /etc/rgy_admin <<! set -h dds://watch state -in_maintenance info quit ! using cron with root as the owner of the process. This resulted in the following: + /etc/rgy_admin set -h dds://watch state -in_maintenance info quit Default object: rgy default host: dds://watch State: in service master replica list is writable Default object: rgy default host: dds://watch State: in service master replica list is writable Unable to go into maintenance state User not authorized to perform operation (network com Default object: rgy default host: dds://watch State: in service master replica list is writable I also made a copy of rgy_admin and modified it's acl's to use setuid root. But this gave the same result. The only way I've been able to get the registry into maintenance mode is to su to root, execute rgy_admin, and issue the state -in_maintenance command but this, of course, won't do. Does anyone have any idea what else I might try? Thanks .../Paul phcalamai@watfun.waterloo.edu
marmen@bcara128.bnr.ca (Rob Marmen 1532773) (09/11/90)
In article <1990Sep10.151351.20186@water.waterloo.edu>, phcalamai@water.waterloo.edu (Paul H. Calamai) writes: > > HELP!!!! There something that I'm attempting with > backups that I'm still having problems with. Try a small c program which executes a SYSTEM call to run the script and setuid the c program. Apollo does not let you run setuid scripts. Note that this setup is insecure. One question: Why do you need to backup the registry? If you have multiple servers and the master goes down hard ( and it does ;-) ), then you can make another slave a master server. We use this technique on five separate Apollo networks. rob... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- | Robert Marmen marmen@bnr.ca OR | | Bell Northern Research marmen%bnr.ca@cunyvm.cuny.edu | | (613) 763-8244 My opinions are my own, not BNRs |
zeleznik@cs.utah.edu (Mike Zeleznik) (09/12/90)
In article <1990Sep11.124044.11537@bnrgate.bnr.ca> marmen@bcara128.bnr.ca (Rob Marmen 1532773) writes: >In article <1990Sep10.151351.20186@water.waterloo.edu>, phcalamai@water.waterloo.edu (Paul H. Calamai) writes: >> >> [unable to put rgyd into service mode automatically; e.g. from a suid >> shell script or program] > >Try a small c program which executes a SYSTEM call to run the script and >setuid the c program. I believe I tried this, and it does not work either. The problem, as I remember it, is with the way rgyd insists you be authenticated (something about password authentication having had to previously occur for this process for the current effective userid... it's been a while, so this could be way off). To my knowledge (someone PLEASE correct me if I am wrong, as this would really simplify my life), there is no way to do this from any suid program, or even from a program which runs directly at login, unless you log directly into the node. Mike Michael Zeleznik Computer Science Dept. University of Utah zeleznik@cs.utah.edu Salt Lake City, UT 84112 (801) 581-5617