in reply to
Re: OT - Unix Administration - What's needed?
in thread OT - Unix Administration - What's needed?
> Be prepared to work long and odd hours (the
> mailserver always goes down at 5 PM on Friday
> night, or 3 AM Monday morning). Your pager will
> go off all the time. Forget your social live -
> you'll see the inside of the server room most of
> the time you are awake - sometimes even 24 hours a
While this is probably the case for many unfortunate sysadmins, I happen to be one of the lucky few with an 8-5 UNIX (well, Linux) sysadmin job. But I suppose what it comes down to is, who you are sysadmining for?
The system I keep up and running is for users who are only in the building between 8 and 5 on the weekdays. Almost none of them ever works late (and no, it's not government work ;-), and weekend work is almost nil. The 2.5 years I've been here, the only weekend I came in to work was for a few hours one Saturday morning for Y2K testing.
Also, we don't host any in-house apps that customers would need to view from the web, we are currently outsourcing that portion of our business. So if that goes down, some unfortunate person at the local ISP will have to handle that work. However, that may not last as our needs for more flexibility grow.
We are also lucky here in that the users don't typically push the limits. On the weekends, people don't want to think about work. So they have no desire to check their email remotely. With that in mind -- if by chance the mail server does go down, our ISP is nice enough to queue all our mail for us until we fix up the server Monday morning. Nobody will miss it while it's down.
Our management is kind enough to respect the difference between work time and personal time. However, I can't say that on some weekends I don't get curious and ssh in to the servers to see what all is going on. If you are looking to save yourself some grief, by all means make sure you set up some tools to monitor the systems, to both watch over your services, and to keep an eye on security. The tools "Big Brother" and "Log Check" are good for that sort of thing.
I have to say though, the thing that really helps keep management off our backs here is the fact that these Linux boxes just never seem to go down. Our DNS servers have uptimes over half a year, and the Linux router we use has over a 400 day uptime. So I tend to sleep very well at night :-)
Does this mean everything is always peachy at where I work? Heavens no. While the boxes stay up, someone has to make sure the services running on them are still working properly. Now, Big Brother helps here, but there's just some things it won't find. We once had an extremely odd issue with one of our DNS servers. The service still allowed you to connect to it, but it was apparently giving back bogus information. Now, I'll tell you, sendmail just didn't like that at all. So instead of quietly accepting that bogus information, or not accepting emails at all, do you know what it decided to do? It decided to take every email that came into our system, and duplicate it hundreds of times. Now, thats all good, but of course this happens on Saturday... and the day after my boss signed up on a high traffic mailing list... and it was one of the weekends where I did NOT happen to check up on things. So, needless to say, it was a messy Monday morning :-)
In any case though, you'll have alot of fun, not much beats a good UNIX/Linux sysadmin job.