Routine problems

[Ram Kumar_1]

1) What are the problems usually SA deals mostly
I know Disk failure is one of them

Any thing else i should be aware of

Thanks in advance,
Ram

[John Bolene · Accepted solution]

Disk failures are few and far between with the current disk drives.

The most of what I do is to add users and change their passwords when they forget them.

Also to clean up file systems and manage disk space when the user puts out thousands of page reports or needs more space for Oracle.

[Steven E. Protter]

attached is a checkhardware script.

It looks for lbolt messages which inevitably lead to disk replacement.

We burned through a lot of disks, so I had cron start running this job. The actual problem was a bad drive cage, through which bandwidth and power are directed to internal drives.

P

[Hai Nguyen_1]

Monitor cpu, memory, disk and network utilizations and take appropriate actions.

Hai

[John Payne_2]

I get calls where people say "It's Slow" They don't know what they mean. Performance should be a big part of what you do. Security is another biggie.

Hope it helps

John

[Pete Randall]

1) Monitor the system.
- this includes disk space monitoring, performance monitoring, etc.

2) Maintain the system.
- periodic patching, user/passwd maintenance, etc.

3) Plan the new system.
- the fun part, planning to take advantage of new technologies

Pete

[Chris Wilshaw]

Let's see, there's.....

Printer admin - eg: adding, removing, troubleshooting

User admin - passwords, enviroment changes,

System monitoring - user audits, space checks, CPU/memory checks, performance checks

Hardware issues - diagnosis and in some cases repair (disk swaps, tape swaps, CPU failures etc)

Security monitoring - failed/bogus login attempts, failed su attempts, various other audits.

System installation and admin - use of Ignite, backup and recovery procedures, extending filesystems, kernel modifications etc.

Network admin - routing changes, monitoring

Other support issues - staff training, project planning, scheduling maintenance jobs, system reporting, testing and development

...I know it seems overwhelming, (and that's not everything), but the key is not to try to tackle everything at once, and to stick at it. Once it starts to make sense, it can be quite easy.

[Chuck J]

We usually gt problems with:

* CPU utilisation
* Swap utilisation
* Disk IO
* Disk space issues
* Memory utilisation
* Rogue processes that don't die off properly


Chuck J

[Eugeny Brychkov]

The usual work of SA is to support operating system and applications, watch for events and failures and report them to management/support in time, suggest and configure systems in SPOF-less configuration (disk mirroring, clustering, alternate pathing) to avoid access/data loss, make regular data backups etc.
If this SA and his company have (hardware) support contract, then only all these above apply to SA job - software support.
But if they do not have support, SA becomes more hardware-oriented, needs to be skilled in hardware troubleshooting and fast hardware problem resolution (parts purchasing and replacement)
Eugeny

[H.Merijn Brand (procura]

My work probably ain't a reflection of a standard SA, but here's a list

1. Keep the systems on line
2. Keep the systems up-to-date
3. Keep the (GNU) software up to date
4. Check that all backups are made
5. Check that they are correct
6. Retreive `lost' files for users from backups (up until a year back)
7. Try solve connection problems
8. Help with scripting
9. Save user's (developer) time by scriptin myself (read: have fun with efficient perl coding)
a. Install new complicated software (Oracle and such)
b. Update existing complicated software

Enjoy, have FUN! H.Merijn