Update manager failure

I recently posted about my problems with patch 118855-36. I finally gave up and decided to wipe the test system clean and start from scratch.

<rant> This now has me tearing my hair out. Sorry, I just don't think that bringing a system up from scratch should be this hard. </rant>

<sigh> Sorry, my frustration leve is rather high right now.

Here is the process I have followed in reloading the system from scratch:

1. Boot the 11/06 CD distribution

2. get a shell login

3. use newfs to wipe the root slice c0d0s0 (I have everything on root so I can do RAID-1 easily later)

4. start the 'install new' process

5. keep the existing partitions by reloading the VTOC (root, swap, a partition for the volume state DB, and /export/home)

6. finish the install

7. bring the system up after the install

8. install the Nvidia drivers for Xorg

9. register the system

10. start update manager

11. update manager show two patches -- install them

12. click on 'check for updates'

13. update manager wedges

14. kill and restart update manager

Well, that's it for update manager. I can't seem to get it to run again without wedging.

OK, I now realize the one of the first two updates was an update to update manager itself. I should have exited and restarted update manager (or even rebooted). Regardless, I should not be able to break update manager just by clicking on "check for updates" again no matter what has been installed.

So here is the thrust of this post: I don't want to reload the system again just to reset the update manager. I need to be able to fix the problem. My assumption is that update manager is a front-end for the patch administration system and that I have managed to corrupt the patch database. I would love a pointer to a tutorial that will explain how to perform administation on the patch management system so that I can repair the database and bring everything back into consistency without having to reload.

Secondarily, I want to get away from any GUI-based tools so I can easily manage remote systems from an ssh login. Sure I can run X back to my local display but I have remote client sites with limited bandwidth and they run VoIP. I can't afford to steal so much bandwidth that they start dropping voice packets. (Yes, I am migrating those sites to routers that support traffic shaping so that I can prioritize SIP traffic but even then I will just have less bandwidth.) I am also not going to allow users to log on as root to run update manager locally. (I do want the servers to keep running.)

So, the questions are:

1. What is the best way to manage system updates remotely?

2. How to I recover from a damaged/wedged update manager?

Pointers to existing doc and tools are greatly appreciated. I am not looking for someone to do this for me but to help me do this for myself. Once I prove the concept of replacing Windows servers with Solaris servers I can get money to spend on training but in the mean time I have to go it alone.

Thanks in advance.

Brian Lloyd

[3182 byte] By [brian.lloyda] at [2007-11-26 17:30:49]

Java programming resources: FAQs, tutorials, compiler and browser download sites, documentation, books lists, IDEs, etc.

Update manager failure

Solaris Operating System New Topic

Solaris Operating System Hot Topic

Solaris Operating System

All Classified