From: visageh@mtn.co.ug
Date: Thu Jan 30 2003 - 02:38:54 EST
Hi all
We are busy implementing DRM 8.7P and in a stretched cluster env. We decided
not to DRM the member root, usr, var, swap disks, but rather will write a
script or rdist to keep files updated from site A to B. {Static data}
Current config, stretched cluster 5.1A TCP interconnect. Site A, SAN node1 +
2 boot disks, root, var, and the rest all DRM disks. I have made entries in
/etc/fdmns and added site B boot1, boot2, root, usr and var. All are mounted
"/SiteB_root", "SiteB_root/usr" etc.
We are running enhanced security. One would think due to the same hardware
being used, SiteA node1, SiteB in disaster node1, one can just copy all the
config files. But that is not the case, WWID for boot + Q disks are
different and cluster on site B will not come up with wrong WWID, major/minor
numbers, quorum disk, etc.
So SiteA Q disk unit/lun 101, boot disks are 102,103.
SiteA root, 104, usr 105 and var 106
SiteB Q disk unit/lun 111, boot disks are 112,113.
SiteB root, 114, usr 115 and var 116
Files I know that cant be copied are:
/etc/sysfonfigtab due to different swap space devices, and quorum major minor
disks.
/etc/fdmns/* due to different boot devices
/etc/clu_bdmgr.conf due to boot device
We intend only to keep /etc in sync.
Are there any other files that I should not copy over, keep in sync ?
What about /var.
Or perhaps for safety just /etc/hosts, /etc/passwd, /etc/services, inet.d,
etc.
A list what to include, or exclude would help, or other suggestions.
Thanks
Hannes
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:06 EDT