Ignite Error:

[Richard_115]

I have a system which when I do make_net_recovery, it passes thro the Version check and soon after, it returns with an error " connection to the server:servername failed". I have done an image of this server before but this time around, I cannot. anybody has an idea about this ?

Thanks
Richard

[MarkSyder]

Is it on a different subnet from the ignite server? This has happened to me too - it works one day and not the next. Infuriating.

The only way I know of to get round it is to move the system on to the same subnet as the ignite server or set up an ignite server for each subnet.

Mark Syder (like the drink but spelt different)

[Asif Sharif]

Hi Richard,

Seems network problem, Is it in same subnet, if not then check it first.

Regards,
Asif Sharif

[Richard_115]

Mark & Saif

I have checked the subnet issue. This cannot be a concern because there are other clients that are on a different network all together and still I can be able to ignite them. Also there are several other systems on different subnets. So this to me will not be a concern. The node I am trying to ignite was on the same subnet when I ignited it last year.
Also, the inetd.conf file does not have tftp commented out. Nor is the services file. In a sense, all things look normal to me.

Any thoughts ?

[JASH_2]

Richard,

The subnet won't matter, it will just come up and ask for you to create a make_boot_tape from the ignite server, so you can ignite it from that tape to tell it about the ignite server and use a lan install.

You do say that it "used" to be on the subnet, have things changed for this server. If the ipaddress has changed or the NIC has been altered, the ignite server will see the differences and think it has no permissions to do the make_net_recovery, or mount any filesystems etc.

Ensure all filesystems are properly exported from the ignite server.
Check that you can ping and telnet to the server from the ignite server and vice-versa.

Any further details you can give me would be a big help.

Regards,

JASH

[Richard_115]

The client systen perfectly mounts the server using NFS. It performs the NIC connectivity check, the patch check and the Recovery tools successfully. Soon after it fails. before here, on the server, if I do "bdf" I can see the server mounted on the client. It however fails soon after checking the Recovery tools.

[Richard_115]

ALL
Here is what I am getting when I run Make_net_recovery.

/opt/ignite/bin/make_net_recovery -s

* Creating NFS mount directories for configuration files.

======= 04/12/06 13:50:28 EDT Started /opt/ignite/bin/make_net_recovery. (Wed
Apr 12 13:50:28 EDT 2006)
@(#) Ignite-UX Revision B.5.3.35
@(#) net_recovery (opt) $Revision: 10.624 $

* Testing pax for needed patch
WARNING: The version of pax on the system to be archived needs to
be patched with one of the following patches or newer ones.
If you do not load the pax patch some files will not be
correctly archived (see the Ignite-UX release notes for
details):

HP-UX Patch needed
------- ------------
B.10.20 PHCO_20388
B.11.00 PHCO_29189
B.11.11 PHCO_28414

you can get patches over the web from the patch database
(http://itresourcecenter.hp.com) or directly from the FTP
server (ftp://us-ffs.external.hp.com/hp-ux_patches).

You have 10 seconds to cancel archive creation...
Do you want to cancel? (y or n)
n
NOTE: Detected entries in the defaults file at
/var/opt/ignite/recovery/client_mnt/0x100090975607/recovery/defaults.

* Checking Versions of Recovery Tools
ERROR: Connnection to the server:pse08 failed.

======= 04/12/06 13:51:03 EDT make_net_recovery completed unsuccessfully

[JASH_2]

Richard,

The obvious question is, have you added the patch requirement to the server you are trying to ignite? When you choose which patch you need to add, it will probably have been superceded by a new one, so just go for the most upto date one that is recommended by hp.

Once added to the server to be ignited, it should create the ignite image ok.

Regards,

JASH

[JASH_2]

Richard,

Looking again at the errors you are getting, the bit where it says:-
/var/opt/ignite/recovery/client_mnt/0x100090975607/recovery/defaults already has defaults in it, could be that the mach address it is defaulting to on the server to be ignited has changed. This could be because it is using a different NIC from before.
On the ignite server you could delete all reference to pse08, including the directory under ...../client_mnt for 0x100090975607 and the link file pse08->0x100090975607 and have the ignite server add a new client, giving it the same name (pse08).
This would then re-create the directory for the mach address of pse08, including its defaults directory.
If you decide to do this, you will need to re-export your exports file.

Before you try this, however, it may be worth taking the end part of the export entry for pse08 off. ie:-
/var/opt/ignite/recovery/archives/pse08 -anon=2,access=pse08
to
/var/opt/ignite/recovery/archives/pse-8 -anon=2

Re-export the exports file #exportfs -av
Then retrying the ignite backup.

Hope one of these works, the second is the easier, but I am not sure it will work in your case.

Regards,

JASH

[Patrice Le Guyader]

Demat Richard,

take a look inside the file defaults in the client directory ( /var/opt/ignite/recovery/client_mnt/0x100090975607/recovery/)
You should have a line RECOVERY_LOCATION=:/var/opt/ignite/recovery/archives/pse08 Verify that this line is ok or it should try to connect to a wrong server or by a wrong interface on the good server.

The advise of JASH about updating the path is probably the first thing to do or why not updating for the last version of ignite on server and clients?

Hope this helps
Kenavo
Pat