ONE PROPOSAL AND ONE ENQUIRY: OAT script and Universe DB on HP-UX cluster

[VK2COT]

Hello everyone,

Happy 2006!

a) Firstly, about my Perl script
HP-UX-check-OAT.pl.

The script does what we call a basic Operations Acceptance Testing (OAT).

The script was developed over several hectic
days, so errors (although not planned) might exist.

My goals were:

* Simplicity to do basic OAT for HP-UX servers;
* Portability;
* Standard Perl interperter (no additional
modules);
* Many new features;
* Support for LVM and VxVM;
* No temporary files;
* No repeated runs of similar commands;
* Not to replace more comprehensive debugging
tools but provide a quick summary of server
status;

Basically, if someone asked me to do a quick check of an unknown HP_UX server, I would use the script to gather information that is very relevant to the support teams (like, is there O/S mirroring, does ODE exist on both root disks, free space and inodes in file systems,
is ICMP blocked to the default gateway,
and many, many others)...

Any comments, suggestions, or results of
run on various HP-UX platforms is welcome.

This script is published in Unix Professions
at HP and it will continue to be supported.

Due to my heavy duties, I will add
more comments into the script later on.
I did not work too much on "cosmetics" like
output formatting as well. That is the easy part once I am happy with the script.

b) Now the enquiry...

Are there any good (or bad) experiences
you are willing to share about
Universe database on an HP-UX cluster
(most probably a two-node cluster).

I tried to search various forums at HP
and elsewhere and did not find enough
information to form a reasoanble opinion.

For example, which volume manager worked
better (VxVM or LVM), failover strategy,
tape backup process, performance "tricks",
and so on...

Universe is not a "mainstream" type of database, and I have seen it at very few
sites (under Tru64 and Solaris).

I shall summarise if I get any useful information worth sharing.

"Edited to comply with ITRC Guidleines"

[Mohammad Sanaullah]

Hi, thanks for such a great script that make admins task easy, However while executing this script on single server its working fine where as in a clustered server it hangs while checking the "CHECKING PRIVILEGED ACCOUNT" and its last output is
WARN File //.rhosts exists;
However on the screen its showing checking for kctune -q maxvgs and its not further.

Same script is running OK in other server not in cluster.

Thanks in advance for the solution.
Mohammad Sanaullah

[VK2COT]

Hello,

I think you are using a two-year old script,
which was modified many times since then.

Could you please get the latest version and
let me know:

http://www.circlingcycle.com.au/Unix-sources/HP-UX-check-OAT.pl.txt

I used my script in many HP-UX environments
and it is always good to find a new
server that has something I did not plan for :)

Cheers,

VK2COT

[Mohammad Sanaullah]

Hi
Thanks for the prompt reply!! but I where using the same script. However I have executed the same but issue still persist.
Here is the last output to the script which I have run on the clustered server.
..
..
..

CHECKING PRIVILEGED ACCOUNT
___________________________
PASS Umask for root set to 022
WARN Root home directory /, not /root
WARN /root directory not valid
WARN Root home directory permissions not 700 (755)
WARN File //.rhosts exists

and running the script which hangs is as follows:
# ./270800.txt >crmdb.oat
lvlnboot: Volume group not activated.
Cannot display volume group "/dev/vg01".
vgdisplay: Volume group not activated.
vgdisplay: Cannot display volume group "/dev/vg01".
kmtune is a wrapper script which exists for compatibility reasons only.
The underlying command used is 'kctune'. New or modified scripts or
procedures should use kctune directly.
About to execute: kctune -q maxvgs


AND after that no responce to the script..
However it works OK to the server not in cluster.
Mohammad Sanaullah

[VK2COT]

Hello,

I see what is happening. The script runs
the following commands:

kctune mqxvgs (if HP-UX 11.20 to 11.23)
kmtune -q maxvgs (if HP-UX 11.11)

For HP-UX 11.31, it checks what LVM VGs are used:

VG version 1, max VGs is 256
VG version 2.0, max VGs is 512
VG version 2.1, max VGS is 2048

So, here is your error analysis:

# kmtune -q maxvgs
kmtune is a wrapper script which exists for compatibility reasons only.
The underlying command used is 'kctune'. New or modified scripts or
procedures should use kctune directly.
About to execute: kctune -q maxvgs
Tunable Value Expression
maxvgs 200 200

# kmtune -q maxvgs 2>/dev/null
Tunable Value Expression
maxvgs 200 200

I do not understand why the script runs
kmtune when your version of HP-UX is 11.23
(or am I wrong)?

I have updated the script for this
interesting problem. TRy it.

By the way, out of interest, what do these commands say on your server:

uname -a
lvlnboot -v
setboot -v

Or, even better, run my script with flag "-b"
and post it to me here:

HP-UX-check-OAT.pl -b

Cheers,

VK2COT

[VK2COT]

One more thing.

Are you using Data Protector for backups?
Commands like devbra(1) can take a long
time...

When you run the Perl script, open another
Shell session and check which commands
run at a time...

Cheers,

VK2COT

[Mohammad Sanaullah]

NO I do not use any Data protector, however pl find the required informations.
UNAUTHORISED ACCESS IS PROHIBITED
HP MANAGED CRITICAL AND CONFIDENTIAL ASSET
$ su - root
Password:
# uname -a
HP-UX crmdb B.11.23 U 9000/800 3021631172 unlimited-user license
# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
/dev/dsk/c2t0d0 (0/1/1/0.0.0) -- Boot Disk
/dev/dsk/c2t1d0 (0/1/1/0.1.0) -- Boot Disk
Boot: lvol1 on: /dev/dsk/c2t0d0
/dev/dsk/c2t1d0
Root: lvol3 on: /dev/dsk/c2t0d0
/dev/dsk/c2t1d0
Swap: lvol2 on: /dev/dsk/c2t0d0
/dev/dsk/c2t1d0
Dump: lvol2 on: /dev/dsk/c2t0d0, 0

Current path "/dev/dsk/c10t0d1" is an alternate link, skip.
Current path "/dev/dsk/c10t0d2" is an alternate link, skip.
Current path "/dev/dsk/c10t0d3" is an alternate link, skip.
Current path "/dev/dsk/c10t0d4" is an alternate link, skip.
Current path "/dev/dsk/c12t0d1" is an alternate link, skip.
Current path "/dev/dsk/c12t0d2" is an alternate link, skip.
Current path "/dev/dsk/c12t0d3" is an alternate link, skip.
Current path "/dev/dsk/c12t0d4" is an alternate link, skip.
Current path "/dev/dsk/c8t0d1" is an alternate link, skip.
Current path "/dev/dsk/c8t0d2" is an alternate link, skip.
Current path "/dev/dsk/c8t0d3" is an alternate link, skip.
Current path "/dev/dsk/c8t0d4" is an alternate link, skip.
lvlnboot: Volume group not activated.
Cannot display volume group "/dev/vg02".
# setboot -v
Primary bootpath : 0/1/1/0.0.0
Alternate bootpath : 0/1/1/0.1.0

Autoboot is ON (enabled)
Autosearch is ON (enabled)

TEST CURRENT SUPPORTED DEFAULT NEXT BOOT
---- ------- --------- ------- ---------
all on yes on on
SELFTESTS on yes on on
early_cpu on yes on on
late_cpu on yes on on
FASTBOOT on yes on on
full_memory on yes on on
PDH on yes on on
CEC on yes on on
#
If there is any issue with the configurations, do let me know this servers is recently configured and made into production.

Regards.
Mohammad Sanaullah

[VK2COT]

Hello,

Even without having full results I see one issue.

Your boot disks are on the same controller.
Normally, I avoid such designs. Luckily for
you, your server is part of a cluster.

Did you try the latest version of the script?

If you wish, you can send me an email to
my HP address (see my details in the Forum).

When the script seems to be hung" ,
you can run commands like ps(1m) or w(1m)
in another Shell and see what it is trying to execute.

I ran the script on clusters many times and
never had a problem, so I am not sure
if and what your problem might be.
But, we can sort it out :)

Cheers,

VK2COT

[VK2COT]

I got no responses and this thread needs
to be closed :)

It was meant as a notification message not
a discussion.

Cheers,

VK2COT