Home / server / Questions

Questions[ndmp](server)

Martin Hope
Sandra
Asked: 2012-06-27 12:44:17 +0800 CST
  • 0

I like to try Amanda's and Bacula's NDMP backup features, but I don't wan't to try it out on the live NAS, just yet.

Ideally would I like to find out which that suites me best, and get familiar with before trying to make a real NDMP backup from the NAS.

Question

Is it somehow possible to simulate a NDMP backup with some Linux hosts?

Or formulated in another way:

Does there exist a NDMP daemon I can install on a Linux host, so it pretends to be a NAS?

backup linux bacula ndmp amanda
Martin Hope
waszkiewicz
Asked: 2011-02-10 07:06:29 +0800 CST
  • 0

I use a NetApp FAS3240 as a storage system, and I'd like to attach an LTO-5 NDMP backup system (with Backbone) in Fiber Channel. What will be the difference of using 1 LTO-5 or 2 LTO-5 streams? Will it improve performance a lot? Is there any kind of LTO-5 technical database with some information?

Thanks a lot in advance for your help.

EDIT: The amount of data could reach 30TB per backup job, no tape backup used yet.

backup tape ndmp
Martin Hope
uuwe
Asked: 2011-01-18 16:22:14 +0800 CST
  • 1

I'm having some issues with a new Backup Exec 2010 R2 installation. I configured a NetApp FAS2020 as an NDMP device and want to backup files from the SAN to a tape drive connected to my backup server. I set up ndmpd according to this document (http://www.symantec.com/business/support/index?page=content&id=TECH48957) and created a separate backup user (http://filers.blogspot.com/2006/09/setting-veritas-netbackup-with-non.html).

Backup works perfectly, but restoring any file gives me an authentication failed error. The NDMP device has a "global" ndmp user configured in the device tab (tried this with the newly created ndmpd backup user and the netapp root) and I can also configure separate resource credentials in the BE restore job.

I have tried setting the same accounts for the "global" ndmp device and the restore credentials and have also tried setting different accounts for them.

NDMP debug level is at 5 and this is what shows up in /etc/messages. The session is closed immediately after it has been granted.

16:12:07 PST [Java_Thread:info]: ndmpdserver: ndmpd.access allowed for version = 4, sessionId = 51, from src ip = 192.168.11.17, dst ip = FAS2020-1/192.168.11.75, src port = 50857, dst port = 10000 16:12:07 PST [Java_Thread:info]: Ndmpd51: ndmpd session closed successfully for version = 4, sessionId = 51, from src ip = 192.168.11.17, dst ip = FAS2020-1/192.168.11.75, src port = 50857, dst port = 10000

Running wireshark on the backup server doesn't produce much. It shows a SYN -> SYN/ACK -> NDMP CONNECT_CLOSE Request from the backup server.

The Resource Credentials for the restore job behave very oddly. If I enter NDMP credentials and do "Test All" it fails. If I use my regular domain backup account, it is successful. There are no failed or succeeded logons in the NetApp ndmp log and tracing this check shows that it doesn't even connect to the SAN. This makes me think that this is more likely flaky BE behaviour rather than misconfiguration of the SAN.

Here is the options ndmp output:

FAS2020-1> options ndmp

ndmpd.access all

ndmpd.authtype challenge

ndmpd.connectlog.enabled on

ndmpd.enable on

ndmpd.ignore_ctime.enabled off

ndmpd.offset_map.enable on

ndmpd.password_length 16

ndmpd.preferred_interface disable

ndmpd.tcpnodelay.enable off

netapp backupexec ndmp