Topic: Server lockup

[Serge Dutremble]

I have a strange situation:

Sometimes, my SME 5.1.2 server locks up for no apparent reason and there is nothing in any log I can look at that gives me a hint as to what is going on.

My system:

DURON 1GHz, 36 MB SDRAM, ECS K7S5A Motherboard, 2 NICs running SME 5.1.2 U1 in Gateway/server mode on Bell Sympatico HSE Residential DSL.

SME was upgraded from 5.0 using the upgrade process using a SME 5.1.2 CR boot.

Also, the system monitor RPM was installed but never starts at boot.  I have to run  the startup script manually from a command prompt.

I also have the annoying PERL warning about the locale after I updated to IMP 3.0.

EasyDNS RPM installed and working.

The server locking up means:

- Internet is unavailable
- Can not SSH into the server
- the server-manager is not responsive
- The console shows nothing special and still shows the login prompt
- any attempt at login returns to the login prompt (after a clear screen) without asking for a password
- A ctrl-alt-del at the console seems to start a reboot but it stops after 5-6 lines with the mention that there are no more processes at this run level.  Any other attempt at ctrl-alt-del only shows a slight disk activity...  nothing else  The system NEVER reboots.
- A cold reset goes through the "disk was not unmounted properly" routine and takes forever (I have one 10 GB and one 60 GB drives).
- Logs show nothing.  All is fine, then a reboot starts...  No shurdown information is shown.

Has anyone ever encountered something like this? I don't know what to think anymore!

This happens about once a week at irregular intervals and has happened once 3 times on the same day!  It seems to be unrelated to my DSL IP change schedule.

Serge.

[Case Jordan]

Been having the same problem lately with SME Sever 5 and the blades.  I don't know whats going on??
My system is a  2x933 MHZ Pentium III, 512 MB RAM, 45 GB Western Dig HD, Soyo Mobo, 2nics dedicated using comcast cable.

It's not a cooling problem, because everything is cool as an ice cube in there.

Same system never locks up under any of the 4.x versions of e-smith.

anyone got any suggestions?
Case

[Case Jordan]

btw, all logs show no problems, no startup problems,
just freezes about every 3rd or so day.

Case

[Chuck]

I think there are bugs in SME 5.1.2 that make it unstable for production machines.  Go back to 4.  It's more reliable.

[Tom Keiser]

A more likely possibility is that your machine bios is not up to date, or you have a subtle hardware problem. There must be thousands of SME 5.x servers running without mishap -- I have, or have set up, a dozen of them. Chuck wrote:
>
> I think there are bugs in SME 5.1.2 that make it unstable for
> production machines.  Go back to 4.  It's more reliable.

[Andrei Taylor]

Sounds like a hardware issue to me. I have installed SMEv5 5.1.2 a few dozen times and it is ROCK stable. I agree with Tom start at the system bios.