Example of hp-ux project.

[Steve TZ]

Hello,

I'am looking for ideas, examples of small hp-ux projects which I can do as junior hp-ux administator. Please don't hesitate if you have some ideas?

Thanks in advance.

[Michael Steele_2 · Accepted solution]

Are you standalone or in a network with other sand boxes to play with?

a) Setup NFS to localhost unless you have another box to act as an NFS client

b) Setup nameserver is a good project, see 'hosts_to_named' command and read about configuring a DNS / nameserver.

c) Sendmail is important to configure but you'll need the name of a Outlook mail server. Note: If you decide to setup your box for receiving emails the be aware that it is a huge security leak. Typically, outgoing is as far as sendmail admins go.

d) Get a system performance set of scripts up and running. If you don't have one then search the ITRC for examples. SEP often publishes his set.

e) Adding in 2ndary swap just for the lesson.

Go to the Freeware HP-UX website and download and install commands like 'lsof' and 'tcpdump'.

[Rita C Workman]

Depending on how much hardware you have to play with:

Set up CIFS/Samba to share UNIX f/s with another UNIX box, or Windows

Set up a 2 node MC/SG cluster

Set up an SSH connection between two UNIX boxes so you can do passwordless,secure,encrypted file transfer.
Then set up SSH connection between your UNIX box and a Windows box.

Now, some of this may be a bit beyond a junior admin...but hey getting these things mastered is what takes you to the next level.

Have fun....and so refreshing to see someone who actually wants to learn it and do themself!! Instead of just saying ..."...someone give me all the answers cause I'm to "xxxx" to even bother to do a search"

Hats off to you Steve!
Regards,
Rita

[Rita C Workman]

Project 1 - Basic Ignite Recovery from tape

Make a recovery tape using Ignite (make_tape_recovery)
See if you can rebuild your box using your new Ignite recovery tape.

Project 2 - Learning the kernel/Performance and Tuning

Do a little research on O/S version and parms.
You already have a tape recovery of your current O/S-so you can go rebuild if you need too!!

Play with the kernel.
Take a view of how your system is running.
Take a list of what the parms are set at.
Learn and use the following:
sar commands (man sar)
vmstat command (man vmstat)
swapinfo command (man swapinfo)
glance
top

Now one at a time change a parm and go back and see what affect that had your system performance.
Example:
Depending on O/S version:

dbc_max_% set to default 50
dbc_min_% also leave at default
(On 11.31 the above would be filecache_max;filecache_min)

nproc
maxdsiz
ninode
maxusers
maxuprc
shmmax
vx_ninode

See not only how the parm affects performance, but take note of when you change some parms, how other parms get changed automatically....which ones...how is box affected. Is the default or auto-changes good?/bad? effect on performance.

Figure out which other ones you might like to check out.

Have fun....
Regards,
Rita

[Steve TZ]

Hello,

Thank you for your feedback. Find my comment below:

Michael

Are you standalone or in a network with other sand boxes to play with?

=> I am not standalone. There are some test servers on the network

b) Setup nameserver is a good project

=> Setup nameserver could have impact on existing environment since we already have dns running.
By the way, I have already tried to setup a dns server (bind) on linux (ubuntu) at home on my computers. It worked. I found it interesting.

c) Sendmail is important to configure

=>Sendmail is already configured.

d) Get a system performance set of scripts up and running. If you don't have one then search the ITRC for examples. SEP often publishes his set.

=> This is very interesting because I am learning scripting. It can also be a benefit for our existing environment.But I didn't understand: SEP often pubishes his set. What is SEP?

e) Adding in 2ndary swap just for the lesson.

=> I have already done this

Go to the Freeware HP-UX website and download and install commands like 'lsof' and 'tcpdump'.

=> I have also already done this. I use sometimes tcpdump but I have to admit that it is difficult to understand the output of tcpdump


Rita

Set up CIFS/Samba to share UNIX f/s with another UNIX box, or Windows

=> This is intersting product. Could this replace NFS if it is used between unix boxes.Apparently it is more secure.

Set up a 2 node MC/SG cluster:

=> Beyond a junior admin I guess :)

Set up an SSH connection between two UNIX boxes

=> I already did this between unix boxes


Project 1 - Basic Ignite Recovery from tape

=> I already did this after a new server installation

Project 2 - Learning the kernel/Performance and Tuning

=> I will do it.

[Benoy Daniel]

Steve,

This may help you.

1. LVM - Boot mirroring, create different types of stripping, Move to LVM2, Modify VG parameters, VG version migration, Dynamic Root Disk.

2. Create Npar/Vpar.

3. Create Ignite image create through network and recover the system.

4. Create software depots.

[Steve TZ]

Hi Daniel,

Thanks for your feedback

1. Dynamic Root Disk seems to be a nice product which gives the ability to clone an HP-UX system image to an inactive disk.

My problem is that on the new HP servers I have seen till now, the disks have sas controller configured. By the way, how can I monitor disks mirrored on hardware level? If a disk fails then I can know. I have seen that ioscan shows one disk when sas is configured.
Example ioscan: disk 1 0/4/1/0.0.0.1.0 sdisk CLAIMED DEVICE HP IR Volume.

Is it possible to unconfigure the sas controller and use LVM MirrorDisk/UX? What is the best, sas controller or LVM MirrorDisk/UX?

3. Create Ignite image create through network and recover the system.
=> This exists already

4. Create software depots.
=> => This exist already

[Stephan.]

=> This is very interesting because I am learning scripting. It can also be a benefit for our existing environment.But I didn't understand: SEP often pubishes his set. What is SEP?

SEP is the acronym for Steven E. Protter, one of the three 2-star Olympians around in the forums.

He put a lot of useful stuff from the itrc on his webpage: http://www.hpux.ws

Or short, i always would pull off my hat for him.

For the projects, i'm pretty sure if you go a little bit in depth Ignite, Performance and DRD will keep you busy for some months.

Another suggestion would be CIMServer, regarding Remote Support more and more important.

Best,
Stephan

[Dennis Handly]

>Stephan: SEP is the acronym for Steven E. Protter

(Acronyms for people's names are generally called initials. :-)

[Stephan.]

My fault, as non native speaker, there is always a word that's missing and I was too lazy to check it.

At least, the initials are an acronym and I did not say something funny. :o)

Best,
Stephan

[Steve TZ]

Hi Stephan. Thanks for the link. It is very intersting site.

If I try to summarize the suggestions from all the feedback, I would choose:

=> Get a system performance set of scripts up and running
I don't really understand what to do. Do I need to write scripts from the begin to the end? Are there scripts which can be adapted or used as example? I will check the site "http://www.hpux.ws"

=> Set up CIFS/Samba to share UNIX f/s with another UNIX box
I will go through the documentation and try this product.

=> Learning the kernel/Performance and Tuning.
I have already changed the kernel parameters but I have never checked the impact on the system. That is a nice exercise

=> depth Ignite, Performance and DRD
I will read the documentation but I don't understand what performance means for Ignite.

[Steve TZ]

Hi Stephan,

Sorry I just did mistake. I wanted to assign points to your first post.

[Steven E. Protter]

Shalom,

http://hpux.ws/system.perf.sh

That is the script set Rita refers to.

There are some fairly interesting things to do on http://hpux.ws

I would definitely go through the install process a few times and get familiar with creating systems, then make ignite backups and practice a few restores.

Cheers.

SEP

[Benoy Daniel]

Steve,

You can't delete IR volume. Instead you can create two IR Volume and mirror it.

[Steve TZ]

Steven,

Thanks for the script set.

Daniel,

So I will keep the IR. But till now I didn't find a way to monitor the disk in case of failure.

[chris huys_4]

Hi,

I think you can delete the IR volume. Else the offline drvcfg utility, would not make much sense.

"
HP 8 Internal Port SAS Controller and HP
Multi-Port Internal SAS Controller Support
Guide
HP-UX 11i v2, 11i v3
"

http://docs.hp.com/en/5900-0112/5900-0112.pdf

Greetz,
Chris

[Steve TZ]

Hi Chris,

Thanks for the useful document.

I found the answer to my previous question "How can I monitor sas disk?". I have to use sasmgr command.

[Emil Velez]

Load integrity VM on a system and create a virtual machine. You can get the software on a trial basis if you need to. This will give you a virtual machine that you need to install. You do not necessarily need extra disks. Your virtual machine can have files (maybe 15 GB) as disks. When you boot up your VM with 2 disks for example you can do some interesting LVM stuff with it.

Just installing the VM from a ignite server might also be a experience.

[Steve TZ]

Hello,

I have choosen "ignite improvement" as project. Could you please tell me what can be improved when ignite server is already installed? Ignite server is already installed. What can I improve?

Thanks in advance.

Steve.

[Steve TZ]

Hi again,

Any ideas what can be improved for an existing ignite-ux server?

Thanks.

[Benoy Daniel]

Steve,

You can automate the ignite backup process by taking a monthly backup of clients automatically.

[Torsten.]

>> have choosen "ignite improvement" as project. Could you please tell me what can be improved when ignite server is already installed?



This is funny.


However, learn how to modify configurations, e.g. different LVOL sizes and custom bundle installations during restore without interactions from a user.

[TTr]

An interesting exercise is having to change the boot disk hardware address either because you change the scsi ID or because you move the boot disk to another controller. It involves booting into "LVM maintenance mode" and doing some interesting LVM and boot configuration editing. This applies more to the older server with standard SCSI disks and less on the newer ones that use the SAS RAID controllers.