"Stephen A. Cochran" wrote: > Someone pointed TKG's product out a few months ago to me, and I looked at their > site. If you read the FAQ carefully, you'll see that you will need one server > for every type of system you need to restore. So if want BMR capability for both > Solaris, AIX, and NT, then you'll need a Solaris, AIX, and NT server for their > product to work. Each of those servers will need to run the BMR File Server and > BMR Netboot Server.From the sound of it, I worry that if you had some servers > running Solaris 6 and some Solaris 7, that you might need a different File/Boot > server for each of those. The BMR File and Boot Servers do not have to be dedicated servers. They can be doing work while they are sitting there waiting to serve up boot images or satisfy NFS requests. A lot can change in a few months. Now, with the 3.1 release, BMR allows you to create CD's that can be used in place of File Servers and Boot Servers. This also allows you to circumvent the use of bootp, bootparams, tftp, and NFS during a TKG Bare Metal Restore if you desire. > After taking over the TSM admin for our site, I've been looking in to BMR > options. The best I've come up with so far is the following: > > I bought some big bit buckets for inside the TSM (RS6000 h50 AIX) server. > Nothing too fast, just decent. For my AIX servers, I'm running makesysb and > dumping the data onto a NFS share. If you are using TSM, wouldn't it be nice to just back everything up to TSM and then use the data in TSM to restore the machine? BMR allows you to do that. The mksysb's back up data that is also being backed up (more efficiently) by TSM. > For Solaris and Linux, I'm using the > hostdump.sh script from backupcentral.com (run by the Orielly backup book > author) to dump to the NFS share. I manually dump the OS only after the change > log is modified (we keep a change log on the boot system so that any admin that > makes a change logs it in case there are problems). This way it's fairly > automated. I also run it manually before/after patching or upgrading. I also run > SysAudit and Sysinfo every night and dump those locally and to the same NFS > share, so I have up to date info on the system in case I loose the disk > configurations. TKG's BMR will capture the machine configuration using a fully automated process. > To restore, I use the install CD to boot, get the networking up and running, > mount the NFS share, and restore. > > I'm still rolling this out, but so far it's worked well. I restored a solaris 7 > server as a test, and it was flawless. If you use Veritas Volume Manager on Solaris, how would you restore the machine with this methodology? It is a very manual process, but BMR handles it with no extra steps or reboots.
begin:vcard n:Schafer;Ray tel;cell:+1 (512) 426-3188 tel;fax:+1 (512) 433-3200 tel;work:+1 (512) 433-3345 x-mozilla-html:FALSE url:http://www.tkg.com org:The Kernel Group;Sales and Services version:2.1 email;internet:[EMAIL PROTECTED] title:Senior Sales Engineer adr;quoted-printable:;;1250 South Capital of Texas Highway=0D=0ABldg 3, Ste. 601;Austin;TX;78746;United States fn:Ray Schafer end:vcard
