| LinHES Forums http://forum.linhes.org/ |
|
| Upgrading multiple systems at once http://forum.linhes.org/viewtopic.php?f=1&t=19533 |
Page 1 of 1 |
| Author: | christ [ Mon Feb 16, 2009 11:17 am ] |
| Post subject: | Upgrading multiple systems at once |
In reviewing the procedure to upgrade to R5.5 from R5F27 I am thinking about how I would upgrade my system which has 1MBE, 1SBE, 3 Front ends. Would I be better off doing: - Auto upgrade of MBE (which has database) - Fresh install of all Front ends using the same system name as I had before Or is there a reason to do an upgrade on the front end? Note 2 front ends will be upgrading with their existing HW and I already built scripts to re-install the mods I have (e.g. fstab, alsa, xorg, etc). One of the Front ends is a complete replacement of an older piece of hardware. I'm really just looking for the simplest path. I'm also concerned that if I auto-upgrade a front end only system that it may muck up the database in some way. Any input is appreciated. C |
|
| Author: | christ [ Tue Feb 24, 2009 11:11 pm ] |
| Post subject: | |
While I didn't get a response to this, I thought it was worthwhile to let others know how I did this. I have my KM Master Back end running on VMware on one of my main network servers as a Guest application. I run a generic myth on CentOS4 Slave backend on the host of the vmware environment so that it can access my PVR500 card. I have three front ends running KM. I did an auto-upgrade on the Master backend. The beauty of doing it in a VM was I could take a snapshot and practice the upgrade. This was a good thing as I needed to do some tweaking and minor work arounds. I could always go back to my working version if things went bad. I then had spare drives for the front ends and did fresh installs of 5.5 on them. I used spare drives so I could test and configure without taking my whole myth network down and thus suffer the wrath of my family. Once I was happy with the results, I updated the vmware MBE, swapped in the new hard drives to all the front ends. Finally I did an update on the slave backend to bring the recorders back in to the network. after the network upgrade, I went around and fixed all the little quirks that happen in a set up this complex. Total actual network outage was less than an hour once I started the network upgrade. It is also worth noting that when I originally set up each machine, I made careful notes about all deviations and work arounds required. I also kept copies of the main files (e.g. xorg.conf, lircrc, fstab, etc). This was handy when I was updating. Christ. |
|
| Page 1 of 1 | All times are UTC - 6 hours |
| Powered by phpBB® Forum Software © phpBB Group http://www.phpbb.com/ |
|