Preventive Maintenance for HP-UX

[Jaweed]

Hi All,

Need UR expertise to show me how to carry out the following activity as Preventive maintenance for HP-UX.

1)Environment Variables ( Temp.)
2)I/P Voltage to workstation – 230V
3)O/P Voltage of Power Supply +/- 5V
4)Fan Condition/Air Flow Adjustment.
5)Checking of Cards / Cables/Connectors.
6)Checking Disk Space utilization
7)Checking system logs
8)Checking the PDC logs
9)OS backup
10)Truncation of old log files
11)Cleaning up /tmp, /var/tmp
12)Search and Removal of core files.
13)Checking the functioning of cron jobs.
14)Verifying backups.
15)Security Check for Unauthorized System Access.
16)Patch Management.

Let me know if I need to check something also as PM.

[VK2COT]

Hello,

You can do these (any many more) manually,
or, if you are like me, automate it.

I was born in the country that some claim
have the laziest people in the world :)
Saying goes that we need a chair next to our
beds to sit and rest after waking up (even
sleeping is hard work for us).

Therefore, I like to automate tasks.
For HP-UX, Solaris, Linux, and NCR MP-RAS,
I wrote Operations Acceptance Testing scripts
in Perl:

http://www.circlingcycle.com.au/Unix-sources/

For example, for HP-UX I run close
to 500 different tests. Not only I summarise
how the server looks like but also if
it works properly.

One of the recent new tests I added was
for /stand/rootconf. Here is part of the
report for an rx7620 server:

AUDIT-PASS: /stand/rootconf non-empty

AUDIT-PASS: /stand/rootconf owned by UID 0

AUDIT-INFO: Hex dump of /stand/rootconf
0000000 dead beef 00a0 0b60 0020 0000 4000 0003
0000010 0000 0000 0000 0000 0000 0000 0000 0000
*
0000028

AUDIT-PASS: Size of file system "/" in /stand/rootconf (2097152 KB) matches
the
size as reported by bdf

AUDIT-INFO: /stand/rootconf seemingly valid ("dead beef" magic label present)

AUDIT-INFO: Possible to boot into LVM

My scripts saved me in many, many projects
when I had to do a quick audit and
nobody was giving me any documentation...

I am sure others in the Forums do it their own way, so you might get more useful
responses too.

Cheers,

VK2COT

[Steven E. Protter]

Shalom,

Items 1-4 Are preformed by you iLo GSP card. they are monitored and all you need to see alerts is to connect once in a while and do SL Show logs.

5 should be done by a visual inspection, monthly is good unless you have a lot of work going on near the systems.

6,7,10,11,12,13,15 should be done by scripts.

Various tricks, see $? output to see if script was successful. There are a number of scripts posted in ITRC to perform these tasks.

9 OS backup should be performed with Ignite make_tape_recovery or make_net_recovery to an NFS share.

Item 14 should be checked with a DR test. You also need a data based backup as Ignite is no good at backing up hot databases.

Item 16 is a manual function of a systems administrator. You should do that after developing a patch policy that works for your organization.

Pretty good list, but its the details and how its attended to that matter.

SEP