Problems with ndmpcopy

Datamover if is different than what we use for backups. Shouldn't any interface be good for ndmpcopy?

server_netstat server_2 -a | grep -i NDMP returns nothing

Telnet to interface works

server_netstat server_2 -a | grep .10000 says *.10000 listening and one other interface is connected to backup machine (it's backing up right now).

--

Jussi

Yes, that should be fine

You should also get an NDMP error in the data mover log when trying to connect via telnet

Try a server_log server_2 | grep –i ndmp - if your last reboot wasn’t that long ago it should also show you nmdp starting and the port number like this:

2010-09-19 14:47:01: NDMP: 6: Listen thread started

2010-09-19 14:47:01: ADMIN: 6: Command succeeded: ndmp port=10000

That sounds pretty generic – see if you are using the correct data mover IP address

Try this - when doing a telnet 10000 you should get back some gibberisch

On the Celerra side a server_netstat server_2 -a |grep -i NDMP should show the port as listening

Rainer

NDMP is up and running on port 10000.

There was no error in log from telnet session to port 10000.

So somehow it isn't listening some interfaces or?

--

Jussi

Hi,

What's your linux ( distribuiton and kernel version ) ?

Are you using the EMC provided ndmpcopy tool that comes with the Tools&Apps CD ?

Please, test with the ndmptool attached to the following solution : https://community.emc.com/docs/DOC-7832

Gustavo Barreto.

Tried also with version mentioned. Original is from tools&apps cd.

Linux is Rhel 5.5. 2.6.18-194.26.1.e15

--

Jussi

I guess then that you either

- have changed the Celerra NDMP settings (port numbers or disabled)

- use the wrong IP

- have something blocking between the Linux client and the Celerra

- have a syntax / usage error in the NDMP command

when I telnet to my data mover and enter some gibberish I do get a NMDP header error shortly after:

[nasadmin@nsdemo ~]$ telnet 10.64.240.124 10000
Trying 10.64.240.124...
Connected to 10.64.240.124 (10.64.240.124).
Escape character is '^]'.
$ÜPuTTY
help
help

`version
Üp
Y
h

server_log server_2 | grep NDMP

1292426255: NDMP: 3: Thread ndmp001 received NDMP request header with error: 1920166255

Rainer

Nothing comes to server log when trying ndmpcopy command.

--

Jussi

What is the NDMP error you see in server_log?

Have you tried without specifying a directory name (profiles) within the file system?

See also the following "how to" on NDMPCOPY https://community.emc.com/docs/DOC-7832

Claude

Ok

I found NDMP error from server logs. Still ndmpcopy won't work.

ndmpcopy command is:

ndmpcopy sourceip:/root_vdm_2/jemma_fs/profiles destip:/jemma_fs_test/profiles -sa ndmpcopy:ndmpcopy -da ndmpcopy:ndmpcopy -level 0

Can't see what is wrong with command.

--

Jussi

Have you created an NDMP user account? I need to know where you are stuck at. There's so much going on in this thread. I'm using NDMPcopy as well and we had problems before, but we figured it out and it's working over here. Our only problem is that level 1 is taking to long to restore. I have steps I can give you that may assist

I am having the same issue in ndmpcopy is there any solution

Which of the test/debug recommendations in the thread did you do and what was the result ?

Can you be more specific?

I am getting error "error:comm.c415 system call (connect):connection refused could not open ndmp connection to host

the command us is ./ndmpcopy 157.146.160.213:/corptst/tpmiipadpdftst 30.25.129.85:/corptst/ -sa ndmpcopy: -da ndmpcopy