When you see this error does the backup fail? Is status of the ssid in media database clear? And if yes, did you try to restore it just to be on the safe side? If that would yes too I would assume this is noisy message which should not get out in the first place and RFE to fix it should be opened with support.

The backup does fail. Here is the CLI output:
================
C:\>nsrnmdsv -f C:\nmd.cfg
01/23/07 13:25:56 nsrnmdsv: Writing debug messages to c:\program files\legato\nsr\applogs\debug-legato
error: task failed for component(s):
HACIENDA_ftindex_01
01/23/07 13:26:50 nsrnmdsv: Error: An error occurred during the NMD task execution: ndmtm_save failed
01/23/07 13:26:58 nsrnmdsv: Task cleanup is successful
================

The debug file is large (D8) but here are the (comparatively) interesting lines (note that I have told it to stop after the first failure):
================
.

.
01/23/07 11:47:08 nsrnmdsv: dump_to_sesslog=0
01/23/07 11:47:08 nsrnmdsv: parse_logfile(): EXIT
01/23/07 11:47:08 nsrnmdsv: task failed: null
01/23/07 11:47:08 nsrnmdsv: error: task failed for component(s):
HACIENDA_ftindex_01
01/23/07 11:47:08 nsrnmdsv: gen_sess_stats(): ENTER
01/23/07 11:47:08 nsrnmdsv: gen_sess_stats(): EXIT
01/23/07 11:47:08 nsrnmdsv: nmdtm_tasks_reap(): EXIT
.

.
path: "D:\\Documentum\\fulltext\\IndexServer";
start: 1169563542;
subdirs: "D:\\Documentum\\fulltext\\IndexServer";
type: FTI;
resource identifier: 0(0);
01/23/07 11:47:13 nsrnmdsv: end of post current resource
01/23/07 11:47:13 nsrnmdsv: nmdtm_sess_report(): EXIT
01/23/07 11:47:13 nsrnmdsv: Number of tasks = 1; Number of successful tasks = 0
01/23/07 11:47:13 nsrnmdsv: warning: All 1 tasks failed
01/23/07 11:47:13 nsrnmdsv: nmdtm_oper_driver(): EXIT
01/23/07 11:47:13 nsrnmdsv: nmdtm_save(): EXIT
01/23/07 11:47:13 nsrnmdsv: FTI backup: TM FTI save job failed
01/23/07 11:47:13 nsrnmdsv: SA unquiesce: Task started. Checking pre-requisites.
.
.
=================

SSIDs are in the media database (but not all that should accompany a NMD scope of ALL). In one place it reported "partial failure" (it seems to fail on an FTI saveset).

R.

Oh yeah, a restore attempt is not possible because the Documentum server is semi-production. I think we can use the mminfo output to see just how "restorable" these saves are (the answer right now is "partially").

Ryan

I'm sorry, but I have no experience with NMD (yet) so I have no idea what to suggest. So I will just think loud and you see if any of it make sense to you:
- if all NMD backups fail chances are that something is not configured correctly
- if it fails "sometimes" is there any pattern to it (like on what does it fail?)
- does it make difference if instead of ALL you do SA_ALL? (full-text index components should exist for ALL to be successful abd 5.3.x comes without it apparently) Testing it with some other values too might help to pinpoint problem. I will also assume you have listed all necessary env. variables in cfg files, but perhaps you wish to recheck (in this case NMD_BACKUP_FTI_QUIESCE, NMD_BACKUP_FTI_UNQUIESCE and NMD_FTI_SUBDIRS).

Above might sound as heavy crap, but then it's just pure guess on something I have never seen and touch. Another thing is that NMD is fairly new module - while I wish not to suggest that it is buggy, I guess some problem might be out there and perhaps that is your problem too. Problem or not, it would good to get support engaged too. If you get down to this let us know what it was.

Hi Hrvoje,

I appreciate your "best effort support". NMD is a new one for me too. I have engaged support twice on this install and they enjoy working with it about as much as I do. The last word actually came from engineering who said that NMD is not supported on ANY cluster...RFE filed.

I am proceeding anyways because we can get around the cluster situation maybe with some pnpc stuff and the AutoStart's scripting ability (or something).

All of that said, I had the same thought as you by altering the scope value. I changed it from ALL to SA and I was able to successfully process all 5 storage areas individually, which was nice. Today I will try to individually backup ICF, DB and eventually work my way back to ALL. My thought that is if the I can get all of them to backup individually, I should stand a better chance at getting ALL to backup.

Ryan

Update:

I'm able to backup the ALL scope and all others with the exception of one FTI object. Not sure what the problem is there but if I try to recover from the ALL backup I can do it just fine (and that's good enough for me).

Other notes regarding NMD:

There is a related bug with MS SQL (297104) that can stop you from completing INCR saves (this could affect anything that uses NMSQL).

We obtained and installed EMC patch LGTpa91122 to address a problem where the savegrp process completes before all the save.exe processes are done writing to tape.

There are 2 known issues with config file variables and how they make INCR saves fail: LGTpa93325 & LGTpa94932.

There are still some problems with this install but things are working (mostly). Thanks again for your help Hrvoje.

Thanks for sharing it Ryan!

Hello Ryan,

We are currently looking into getting Legato NMD and were needing to know some more info about it, such as how it works, where it is installed to, etc. Do you know of any sources of this info? I've tried to search PowerLink but I really can't find anything relevent (and I can't look through all 119,xxx hits).

Thanks!

Go to Home > Support > Documentation/White Paper Library > Software J-O > NetWorker BusinessSuite and PowerSnap Modules and on left side you will see link to NetWorker module for Documentum.

I recommend that you read the documents and setup a lab as well. You will learn a lot as you work on the install and testing. I found it to be a bit of a steep curve but it would have been much easier (and more fun) if I weren't on a production system.

R.