ndmpcopy fail

[iversonshiu]

when i start ndmpcopy -f ,spend some time waiting for the above.

FAS3210B*> ndmpcopy -f /vol/vol0 /vol/vol8

Ndmpcopy: Starting copy [ 4 ] ...

Ndmpcopy: Socket connection to source FAS3210B failed

Ndmpcopy: Issue 'ndmpd on' on the source filer to enable NDMP request then retry the connection

Ndmpcopy: Done

FAS3210B> options ndmp

ndmpd.access                 all       

ndmpd.authtype               challenge 

ndmpd.connectlog.enabled     off       

ndmpd.data_port_range        all       

ndmpd.enable                 on        

ndmpd.ignore_ctime.enabled   off       

ndmpd.offset_map.enable      on        

ndmpd.password_length        16        

ndmpd.preferred_interface    disable    (value might be overwritten in takeover)

ndmpd.tcpnodelay.enable      off       

but , i check the options ndmpd and ndmp status , not any problem...

please to help me , and give me good solutions,tanks

[scottgelb]

There can be issues when the primary interface (check /etc/hosts and see where the hostname FAS3210B is used.. the one with more than one name with both the interface and host names) then make sure that interface is up.  Often we find when it is down ndmp local won't work...even though it uses the loopback adapter it needs the primary interface defined to the host online.

[BB_STEINFELDT]

Hi there,

are there any updates to this issue? We experience the same problem, "Issue 'ndmpd on' on the destination filer to enable NDMP request then retry the connection"

hosts file looks good, interfaces are all up, name resolution works fine.

This issue occurs on a local ndmpcopy job only, from a remote filer the ndmpcopy job runs well.

[BB_STEINFELDT]

We could solve this issue. This is due to a bug in OnTap 8.1 and some others too, but fixes in Release 8.1.1

There are Zombie NDMP Sessions left which prevent to start new ndmp jobs. Here´s the Solution:

OEM: Data ONTAP 8.1 NDMP daemon does not allow new NDMP sessions

Symptoms

Protection manager jobs or NDMP backups fail for a NetApp storage controller running Data ONTAP 8.1.

The following error message is displayed:

ERROR: maximum number of NDMP processes (392) forked.

The ndmpd status command displays no ndmpd sessions.

Cause

In Data ONTAP 8.1, the ndmpd daemon follows the UNIX convention of forking a copy of itself for each NDMP session. For some reason, the daemon gest corrupt and cannot fork any copies, even though there are no copies running.

The following is a BUG in Data ONTAP 8.1:

BUG 587756 - NDMP service disruption, NDMP does not serve any client requests

Solution

In the public BUG report, the workaround says 'Restart NDMP service,' but is sparse on details. The following is the procedure to do that:

The ndmpd daemon needs to be restarted to clear it and allow it to start spawning copies again. Rebooting the storage system will accomplish this, but getting permission to reboot is difficult for many users.

Perform the following steps to stop and restart only the ndmpd daemon process:

1. Open the systemshell
   1. Unlock the 'diag' userid if not done previously
      For more information, see article 1012484: Data ONTAP 8: How to unlock the diag user account
   2. filer> priv set advanced
      filer*> systemshell
      login: diag
      password: <diag userid password>
      filer%
2. Find the current NDMP daemon
   filer% ps aux | grep ndmpd
   The output will look similar to the following:
   root 1693  0.0  0.6 20364  4328  ??  Ss  Fri12PM  0:05.38 /sbin/ndmpd
3. Stop the current daemon:
   filer% sudo killall ndmpd
   Messages similar to the following will be displayed, indicating the daemon process has been terminated:
   Tue Sep 4 13:43:46 EDT [fas2220cl1-rtp:spm.ndmpd.process.exit:error]: Network Data Management Protocol server (ndmpd) with ID 1693 exited as a result of signal 15. This server will attempt to restart.
   
   Tue Sep 4 13:43:46 EDT [fas2220cl1-rtp:spm.ndmpd.procesfas2220cl1-rtp% s.exit:error]: Network Data Management Protocol server (ndmpd) with ID 1693 exited as a result of signal 15. This server will attempt to restart.
4. Determine if the daemon has restarted:
   filer% ps aux | grep ndmpd
   If one /sbin/ndmpd process is displayed, the daemon has restarted normally. If multiple '/sbin/ndmpd' processes are displayed, the daemon has restarted normally and NDMP jobs have started running. The ndmpd status command run at the Data ONTAP prompt should display an NDMP session for each of the 'extra' daemon processes.
5. Exit the systemshell, and reset the privilege mode:
   filer% exit
   filer*> priv set admin
   filer>

Note: BUG 587756 is listed as fixed in Data ONTAP 8.1.1.

[MooseQuest]

Interestingly enough I worked around this issue by specifying localhost as the filer.

ndmpcopy -sa user:pass -da user:pass locahost:/vol/myvol localhost:/vol/myvol2

[Noorain02]

While running ndmp copy between 7mode filer to C-dot getting below error, though configuration is correct. There is no data in source volumes.

Ndmpcopy: vserver_IP: Data: START RECOVER parse failure on nlist: incorrect path specified
Ndmpcopy: vserver_IP: Data: RESTORE IS ABORTED
Ndmpcopy: Failed to start restore on destination
Ndmpcopy: Done