Disaster Recovery on a SUN platform
1995-12-13 11:42:30
This is what we use for Disaster Recovery on the SUN clients. Our SUN folk
wrote this and they have tested it to their satisfaction. We have about 100 SUN
servers and ADSM backups up an average of 7GB per night. My high water mark was
21GB one night. Any feedback will be passed on to the SUN folks for
clarification.
Michael Bryant
Fingerhut Corp.
Disaster Recovery Steps for SUN Workstations.
This disaster process assumes you have backed-up the root
partition on the SUN workstation with a "ufsdump" file image.
This image is then put onto your ADSM server for
archiving on tape. This "ufsdump" of the root partition
gives you the disaster recovery ability because ADSM
does not handle special SUN driver files. These files
are not restored correctly by ADSM. Only "cpio" and
"ufsdump" can only restore special files correctly.
1) Boot SUN workstation from minimal jumpstart kernel or
from cdrom.
2) Configure the interface to work and start-up rpcbind.
3) Add ADSM server host name into /etc/hosts and edit
resolv.conf file to access host vi DNS.
4) format your new hard drive with appropriate partition
spec's.
5) Mount your directory partitions on your new disk drive
for restores.
6) Install ADSM client onto local machine.
7) Run "dsmc" from the command-line with options to restore
each file partition. (i.e. /usr, /var, /export, ...)
To restore the root partition, restore the "ufsdump" image
of root to the local machine. Then "ufsrestore" the root backup
image to the root partition.
change directory to the mounted root on harddrive not active root.
i.e. cd /; ufsrestore -f /tmp/rootdump
8) Install the boot block from cdrom or nfs mountpoint onto
the local machine.
9) reboot machine and check that everything is the way it was
from the last backup you restored from.
10) done.
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- Disaster Recovery on a SUN platform,
Michael Bryant <=
|
|
|