Slashdot Mirror
30 Years a Sysadmin
itwbennett writes: Sandra Henry-Stocker's love affair with Unix started in the early 1980s when she 'was quickly enamored of the command line and how much [she] could get done using pipes and commands like grep.' Back then, she was working on a Zilog minicomputer, a system, she recalls, that was 'about this size of a dorm refrigerator'. Over the intervening years, a lot has changed, not just about the technology, but about the job itself. 'We might be 'just' doing systems administration, but that role has moved heavily into managing security, controlling access to a wide range of resources, analyzing network traffic, scrutinizing log files, and fixing the chinks on our cyber armor,' writes Henry-Stocker. What hasn't changed? Systems administration remains a largely thankless role with little room for career advancement, albeit one that she is quick to note is 'seldom boring' and 'reasonably' well-paid. And while 30 years might not be a world's record, it's pretty far along the bell curve; have you been at it longer?
Oh look a mouth breathing 600lbs 50 year old virgin living in his 80 year old mother's basement...still.
and ghost of Rear Admiral Grace M. Hopper might decide to choke this fool with a microsecond.
Been a System Programmer/System Manager/Sysadmin since 1976. Only worked at three different places...not bad, nicht wahr?
There is no God, and Dirac is his prophet.
Must have been a Zilog System 8000. Probably a model 21, as that was dorm fridge sized. The 31 and 32 were the size of a full size refrigerator.
Ran a Z8000 series processor at 10 MHz, and had about 8MB (if you were lucky) RAM. The hard drives were about 40MB and had an SMD interface.
They ran ZEUS, which was Zilog's System III variant.
I loved ours.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
I've been doing systems work of some kind since the early 90s. The technology changes a lot, but learning the fundamental concepts early on will allow any sysadmin to continue being productive even when entire platforms get swapped out from under you. Unix --> Linux, Windows GUI --> Windows PowerShell, Physical servers --> Virtual servers, Virtual on-site servers --> cloudy virtual servers -- all these transitions can be made successfully by falling back on the fundamental tasks of controlling access, dealing with failures, providing resources, etc. that are similar at their core no matter what you're running on.
The thing that trips up a lot of sysadmins is getting bogged down in the details of one particular platform or aspect of their job and not seeing the big changes that come up. For the right kind of crazy person, this job is actually fun. I hope I'm doing something like it years from now.
That's the modern tension. Old School folks want to understand how things work and the new school just wants to run applications some expert wrote and not worry about the details. The latter is more productive when it works.
But, the former is needed when it doesn't work.
"Be particularly skeptical when presented with evidence confirming what you already believe." -
Since 1991:
System 6/7
MS-DOS
386BSD (no, not BSD/386... 386BSD, Jolitz's version)
Windows 3.11
Solaris
IRIX
AIX
Linux/Windows Server 2012R2
What has changed since 1991:
1: More reliance on networks for security. Most places, if the firewall or core fabric get compromised, everything is hosed. Since Windows usurped Solaris as the primary Internet based OS, it pretty much is assumed hosts will be compromised or misbehave.
2: Less interest in what is happening at other sites. Used to be that if someone at a.com had an attack coming from them, email to postmaster@a.com or a phone call from the InterNIC record stopped it in the tracks (with the user being properly LARTed.) People even had identd so if a misbehaving user was causing trouble, other remote sites would know who it was... which kept IRC sane for a long while.
3: The back-turning on security. Used to be virtually every company took security seriously. Now, unless someone rm -rfs /ifs/data/* on the core Isilon cluster, nobody gives a rat's ass, since security has no ROI.
4: How shitty IT people are treated. Sysadmins used to be treated with some respect; basically the priests of the temple of Syrinx. Now? Viewed as fungible with cheap H-1Bs supposedly able to do what they do, except cheaper and 100% loyal (or they get deported).
5: The offshore fetish. This is sort of equivalent to an enema fetish or vacuum bed fetish, except more messy. Move it offshore, even though it causes major delays and code quality issues with development, the initial costs are cheaper.
6: Pertaining to #5, the fact that code quality has gone to crap. What would be an in-house version never seeing the light of day is now a pre-release candidate. "If it builds, ship it", is the motto now.
7: No interest in backups. SANs are reliable, but it isn't tape, and it is only a matter of time before some hacking group starts purging SANs as a matter of course. Yes, the controllers are behind a management network, but there is always island hopping and unknown bugs. RAID isn't backups, replication isn't backups (it replicates the "rm -rf /"), snapshots are not backups, the only thing that are backups are copying to media that can be stored offline with 0 watts needed (other than HVAC/environmental items.) This already has killed a local company here in Austin when someone knocked their main servers offline.
8: The fact that if you even have a -hint- of being depressed or anxious, you will be fired -stat-. So, you have to always pretend to be 100% "sane". Someone dies in your family and you are bereaved? Grounds for termination. Again, it is an offshoot of #4.