Linux as a server OS?

[mindwarp.out]

I am looking at some stuff over at byte.com (they had the article on BSD vs. Linux off of slashdot), and alot of the other articles show like redhat 7 crashing after just a couple of hours. Do most servers crash so often? I cant imagine my debian box crashing that much,
but I am wondering: what should the average uptime of a high demand server be?

Wouldn't a server crashing every few days require a ton of administrative attention? Just to check that all the services are running? Is it just more intelligent to make like crons that restart the daemons? Is it even a worry?

Mindwarp

[ph34r]

Most servers running a *nix or *bsd are up for weeks, months, and years. Redhat 7 had serious problems, but something like Debian or Slackware should be able to go on for nearly forever.

[mindwarp.out]

I guess I am worrying about detecting when a service goes down and restarting it without admin intervention

Mindwarp

[X_console]

If you're worried about a service going down, then you can simply write a script that checks if it's up. If it's not, the script restarts it. You can then have this script run by cron every few minutes.

My Slack server was up for 4 months before it had to be rebooted for administration purposes. But no, normally it doesn't crash on me.

[Ryeker]

My Slackware server was up for almost 6 months straight. It was put online 6 months ago. Never once went down. Until I decided to 'upgrade' it... hehe... well, that didn't turn out too well (should have been in single user mode...). RedHat has different versions of the distro. Ones like a server version (high availablity or something like that) and a general workstation. That could be the problem.