Re: Tao-discuss Digest, Vol 27, Issue 3

[email protected] wrote:
> Hi !
>
> I've got a remote machine that is some 200 mile away. I've just re-installed
> the machine from scratch a month ago with tao 1.0 up to date. (I need 1.0
> because it's a dell with some raid card not yet fully supported in 4).
>
> Since a couple of week, sometimes, I cannot login remotly. I get the
> password prompt, I enter my password, and it hangs there. ssh -v didn't show
> anything usefull. When that happens, I have to tell the IT guy there to
> reboot the machine. After that it's ok generally until the next day, where
> it starts again to hang.
>
> I suspected a cron running at night or something like that, but the jobs
> running at night runs ok from my shell. I didn't see nothing intelligent in
> any log file.
>
> One day, I stayed logued in running top, and the next day, I saw that the
> uptime was 5 (very anormal for that machine usually having uptime of 0.2 or
> the like). So I suspect a lack of ressource of some sort, but I can't point
> my finger on it.
>
> The machine has plenty of ram (1 gig, 2 gig swap), plenty of scsi disk space
> on all partitions. It's only a web server with mysql and some other things,
> nothing major...
>   

    First thing I'd do is install memtest and have it run locally for 
about 12 hours and see if there are any errors.

    A hang like that sounds to me like something in RAM is getting 
corrupted. Once that has been eliminated, I'd look at the other 
hardware. You could also try uninstalling and reinstalling the SSH 
server, in case the original package was corrupted in some manner.

    Michael