Network protocols behind Ignite?

[HP-UX TIM OSP]

Hello:

What network protocol is Ignite using during an OS depot installation after it NFS mounts the /var/opt/ignite/client directory and pulls the client config? I see allusions to tftp but I can't find a definitive answer.

Thanks,
Nelson

[Steven E. Protter]

tftp can be used to push/pull the install kernel to the client when beginning the process.

The Berkley r- protocols are currently used to gain control of the client when pushing/pulling an image at the beginning.

As you note, NFS is used for transferring the golden image archive to the target machine.

SEP

[HP-UX TIM OSP]

Hi Steven:

How about for OS depot installs, e.g., /var/opt/ignite/depot/Rel_B.11.11/core ? After the Berkley services are used to get the client ready and prepped for installation and the NFS mount of /var/opt/ignite/client occurs, what happens next?

Thanks,
Nelson

[Steven E. Protter · Accepted solution]

Due to a few failed Ignite pushes, I can tell you what happens next.

After connection:
The core os image is transferred. In configuration, you have several choices, including nfs and ftp, maybe tftp.

Then the core OS is "installed"

If its a golden image push, after a few intermediary steps, the golden image is transferred tot he target client via nfs.

Next pax_iux is used to unpack the image and lay it out in a temporary filesystem.

After thats complete something I've not monitored happens to change the tempoarary filesystems to permanent names.

Then post installation steps and user controlled scripting happens.

Did I answer your question this time?

SEP

[HP-UX TIM OSP]

Hi Steven:

You actually might have. Based on the config I saw for the depot and knowing how the OS is installed through SD-UX, it seems to me that the client actually gets a mini-system going then does a remote SD-UX swinstall of the OS depot housed on the Ignite server which I guess behind the scene NFS mounts the appropriate depot somewhere on the client so it does an installation. What do you think?

Thanks,
Nelson

[Steven E. Protter]

Based on experience, I think your conclusions are correct sir.

SEP

[John de Villiers]

fptp is used to get the version of the ignite server.

It uses that to get /opt/ignite/Version to get the version of the ignite server.
It Then uses NFS to mount /var/opt/ignite/clients/
and NFS to mount /var/opt/ignite/recovery/archives/hostname

The second mount can be changed in the gui so its not always the same.

If the version of the file fetched with tftp does not match it will initiate a swinstall from the depot on the ignite server.

John

[John de Villiers]

erk - finger trubble - make that tftp.

-----
We are Dyslexia of Borg - Fusistance is retile, your ass will be laminated.

[Zeev Schultz]

bootp,tftp, to boot (obtain dhcp-like settings) and fetch install kernel.Later
install depot using swinstall (which use rcp I think).

[Elli Wang]

It will depend on your config files.

Your archive can use NFS, FTP, TFTP, RCP.
Ignite log use NFS
Scripts and boot kernel use TFTP.
SD-UX depots use SD-UX.

You said OS depot. I assume it is SD-UX depot. Then it's using SD-UX.

BTW, if you do not use golden images, you will need at least NFS, TFTP and SD-UX.