ISA Server with 2 Network Cards

[getndz]

Hi all,

I've installed an ISA server ( 2 network cards) and i cannot backup it.
Ping is OK from and to commcell server.
No problem to access to the shares in both directions.
Data Interface Pairs is not the solution.
DNS and Host file OK.

Any idea ?

regards

 

[Cyklops]

Firewall restrictions ?

Have you tried setting up a data interface pair by defining both servers by IP-adresses instead of FQDN?

regards

 

[birky]

yes, try DIPs and replace hostname with IP's...should work

 

[Frbutler]

what errors are you getting is it a authentication issue or a network issue?

 

[Frbutler]

also is ping ok from client to media server?

 

[getndz]

Back from the holidays...

To Cyklops: yes , tried with interface pair= no chance.
To Frbutler: this is one of the messges ( "detail" )
Loss of control process ifind.exe. Possible causes:
1. The control process has unexpectedly died. Check dr watson log or core file.
2. The communication to the control process machine ischla03.industrie.ammann.intra might have gone down due to network errors.
3. If the machine ischla03.industrie.ammann.intra is a cluster, it may have failed over.
4. The machine ischla03.industrie.ammann.intra may have rebooted

I can ping from and to the Commcell server, no problem.

Regards
getndz

 

[Frbutler]

Here is the standard CV answer

KB Article ID: 10975
Article Type: Troubleshooting
Published: 6/25/2007 3:15:37 PM

Loss of control process ifind.exe.

Symptom:
File system backup jobs fail and show Dangling Cleanup as their status with the following error:

Loss OF CONTROL PROCESS ifind.exe. Possible causes: 1. The CONTROL PROCESS has unexpectedly died. Check dr watson log or core file. 2. The communication to the CONTROL PROCESS machine may have rebooted.

You may also see the following event message:

Processes for job [] were forcefully terminated.

Cause:
The ifind.exe PROCESS runs in the initial stages OF File System backups to gather a list OF files on the computer. This error indicates that ifind.exe has not responded to JobManager in a timely fashion.

The most common cause for this error is when the client's Event Manager service (EvMgrC.exe) is down or unresponsive.

In other situations the PROCESS may be dying due to a mismatch in binaries or you may see that there are multiple instances OF ifind.exe running.

Resolution:
First, check the PROCESS tab OF Windows Task Manager on the client computer to see if there are multiple copies OF ifind.exe running. If you see multiple copies OF ifind.exe, please select each OF them and click End PROCESS to kill these processes.

Check the client computer's Galaxy services. Client Event Manager service might be down, or if it is still running might be hung. Restart the services on the client.

Resume the job. If the job continues to fail with the same error, the problem may be due to a mismatch in binaries or involve network issues. Please gather the logs for the job in question and contact software support for assitance.

 

[Frbutler]

My money is on mismatch binaries
Is the client on the same version as the media agent?
SP level?

Other than that it is a Service restart or network issue.

You can try removing all updates and reloading SP4.

Let us know...

Frank

 

[getndz]

No ifind.exe running at all on the client server...

Another message ( job "detail" ):
[CVSession::socketConnection]:Remote machine [ischla03.industrie.ammann.intra]. The socket connect failed. Error returned [-1=Timed out while waiting for data in socket.]. Check Network Connectivity.

Services restarted
Ping OK

Still the same problem

 

[Frbutler]

Clear the log files (Copy to another location)
re-run the jobs let it fail and then do the log file shuffle

Delete any DIP you have set for it make sure the client has a host file entry on the media server for the specific nic you want to use for backups

Change the properties of the client to reflect the IP address as the host name... I think CV is getting confused on which nic to use

 

[getndz]

I'll try...

New Message:
Waiting for the services on the client [servername] to come online.

The services are running

 

[Frbutler]

Uninstall the client and reinstall
When it asks you to provide the interface name put the IP address of the nic card in pace provided

Your client is defaulting to FDQN

I have multiple servers with dual nics, This is what I do.

Or.... you can add a host file entry that says
<IP addy> Server_bdn.domain.com

The bdn designates your backup nic
and use that fdqn in you cv client setup (under interface not client name)

 

[Frbutler]

before you do the above make sure you add the client backup nic IP into the host file on you commserve and all media agents

Is this a 6.0 landscape?

 

[getndz]

It's already done...in all the host files.

172.xx.xx.xx Commcell.domainname.com

172.xx.xx.xx Clientname.domainname.com

172.xx.xx.xx MediaAgent.domainname.com

I'll try to reinstall Galaxy on the client server

 

[getndz]

Reinstalled Galaxy , client interface name = IP address

Message:
[CVSession::socketConnection]:Remote machine [172.16.8.30]. The socket connect failed. Error returned [-1=Timed out while waiting for data in socket.]. Check Network Connectivity.

I've to kill someone soon....may be the Commvault staff.

 

[Frbutler]

is this 6.1?
Are all the DIP deleted?

Might be time to do the log file shuffle with support
You can run GXTail.exe from the base directory
Have it watch the cvd.log file (Commserve). then kick off your backup
It will show you real-time what is going on.

 

[getndz]

Is 6.1., SP4 and Post SP4
DIP is deleted.
Host file OK
I've opened a case 1 month ago with the Commvault support, sent the logs, defined the DIP and so on but no result till today.

 

[Frbutler]

did you run GXTail and see where it was hanging?
I would call your sales guy and rip him a new one
Maybe his SE can help or he can expedite the case

I have done that in the past and got results