My setup is quite simplistic. ALL servers are running the same kernel and the same /usr. If I need to update the linux system, I do it on the maintenance server and then propogate it to ALL production servers. This might not work for a complex set of servers (database servers vs webservers), but there you could have parallel maintenace systems.
/Thomas Kern /301-903-2211 > -----Original Message----- > From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On > Behalf Of Levy, Alan > Sent: Thursday, July 08, 2004 12:04 > To: [EMAIL PROTECTED] > Subject: Re: Linux under VM and Cloning > > > We tried making /usr a ro minidisk. It did not work for us. > When we had > to upgrade our kernel using an RPM, it failed since it tried to load > files to the /usr which was not owned by the clone. > > How did you get around this ? > > Alan Levy > W: 718-403-8020 > C: 347-203-0638 > Nextel: 172*26*9628 > > -----Original Message----- > From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of > Kern, Thomas > Sent: Thursday, July 08, 2004 11:51 AM > To: [EMAIL PROTECTED] > Subject: Re: Linux under VM and Cloning > > The production copy of the /usr minidisk is NEVER linked RW > by anyone. I > have a separate maintenance instance where I install and maintain the > linux > system. It has a /usr minidisk of the same size as the production /usr > minidisk. Both minidisks are OWNED by a placeholder named LNXDASD at > virtual > addresses 1000 and 1001. When I am ready to put a new /usr into > production, > I properly shutdown the maintenance user and the first of the > production > servers. I modify the directory entry of the first production user to > use > the new /usr minidisk (ie change LINK LNXDASD 1000 0592 RR to LINK > LNXDASD > 1001 0592 RR). Then I bring up the production server and make > any local > changes necessary like copy files from /usr/upgrade/sbin to /sbin, or > /usr/upgrade/etc to /etc. I test that server and when I am satisfied > that it > works there as well as it did in the maintenance/test server, > I move to > the > next production server and repeat. When ALL production > servers are using > the > new /usr, I DDR copy the production level /usr minidisk (now 1001) to > the > old minidisk (1000) and modify the maintenance user's > directory entry to > use > the 1000 minidisk for /usr and start all over again. Until this point, > my > backout is to switch the server back to the old /usr minidisk, ipl and > fix > local files. > > /Thomas Kern > /301-903-2211 > > > -----Original Message----- > > From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On > > Behalf Of Seader, Cameron > > Sent: Thursday, July 08, 2004 11:27 > > To: [EMAIL PROTECTED] > > Subject: Re: Linux under VM and Cloning > > > > > > How did you setup that read-only /usr minidisk? did you link > > to a minidisk from VM or is it mounted on another guest as > > read-write? > > TIA > > -Cameron > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: INFO > LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: INFO > LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
