system locks up when running /scripts/upcp

sushi3

Registered
Oct 2, 2003
2
0
151
Hey y'all.

Long-winded explanation here. Thanks for taking the time to read this.

I decided today that I'd update cPanel. I've never had a problem doing that, but today, the thing locked up in the middle of the update. I'm running a dedicated server with CentOS 4.4, with 1GB RAM. I have ~650MB free, ~60KB used on the swap partition, so I highly doubt it's a memory issue, but the ISP sysadmin told me there was an "out of memory" error, caused by an HTTP process, on the screen when he rebooted. Strange.

Here's how I got to where I am:

  1. I ran "/scripts/upcp --force" via SSH
  2. It got to the part where it said
    Code:
    cPanel Layer 2 Update Commencing
    ==> Running post install scripts....sync
    ........
    (I believe) with a lot of dotted lines, indicating the process was still continuing.
  3. After several minutes, all my services, including SSH, stopped working, for over an hour.
  4. I had to get my ISP to reboot my server. All the daemons came up fine, except for cPanel's.
  5. cPanel now does not run at all, and it looks like a lot of files are missing.
  6. When I tried to start cpsrvd manually, it said there were things missing.
    First it was "Lchown.so", fixed that by making a symbolic link.
    Then it was "vxs.so", so I made a symbolic link to the "version" directory under the Perl installation.
    Now the error message I get is: "BSD::Resource object version 1.25 does not match $BSD::Resource::VERSION 1.28." which I can't figure out how to solve.
  7. Extreme frustration ensued.

So, any ideas on how I can get upcp working again? (It's never given me problems till now.) I read somewhere to just leave it running, but I don't see how that would solve anything considering the memory consumption effects.

I don't have any customizations except for exim.conf and antivirus.exim. Fairly standard install. I did manually upgrade PHP to 4.4.6 and MySQL to the latest 4.1.x version. I wish I could tell you what version of cPanel I was running. It was probably about 6 months old.

I did file a ticket with my ISP (since I got my license through them). Turns out they charge -- reasonably, but still... I'm wondering if there's a quick solution.

Thanks for any light you can shed on this...

- Sushi
 

sushi3

Registered
Oct 2, 2003
2
0
151
wow! lucky break for me...

Hey guys.

By some fluke of supernature, I had turned on automatic updates to STABLE via WHM, before the fubarration. Just now, the auto-update actually installed cPanel and WHM for me, and now they're both fully functional!!! (Crazy...) I realized this when I stopped seeing my "cpsrvd automagically restarted" emails and read through the /scripts/upcp email.

But... that still doesn't really answer why a "/scripts/upcp --force" via SSH would initially cause such huge problems, with the server becoming completely unresponsive and out of memory.

I guess I should just stay away from SSH updates... ??? I rather liked the thought of updating manually, since I've had problems with auto-updates in the past. And I liked the option of having the --force option since that seems to fix practically any cPanel problem I've read about.

Any thoughts, anyone? (Or am I just talking to the voices in my head?)

Thanks...

- Sushi