nwadmin program responds too late to requests.

[LegLegato]

We have a Networker server version 6.2 whose nwadmin program responds very slow while performing different tasks in there. While ejecting tapes from autochanger this programs freezes for about 20 to 30 seconds before it responds. It accepts a few clicks and freezes back. This is very time consuming for operators while they eject and deposit tapes into changers. Any idea? I am pretty sure that the problem is in media db. I tried nsrim -X and didn't help. I stopped services, renamed mm folder and started back the services. Started nwadmin and worked great. Quick response and didn't freeze. I stopped the services deleted the new mm folder generated and renamed back the old one to mm. Started services started nwadmin program. Situation is the same. Software freezes the same. CPU and memmory utilization on server are less then 10%

Any help is appreciated.

 

[RichS24]

you didn't mention which o/s and if this server has been upgraded.
If you use solaris, and upgraded through version 6.0, seriously expect your media db to be corrupt. Cheers
richs24

 

[TDun]

It does sound as though you have a blitzed index.
What is the speed like if you do the ejects using the nsrjb command from a command prompt?
How many tapes does your media index know about? If not many, then creating a blank media index and scanning the volumes in may be the cure (but tedious).

 

[LegLegato]

I'm sorry. OS is Windows NT sp6a.

Using nsrjb from the command prompt is much better. Problem is with the GUI of nwadmin program. winadmin program runs without a problem and does not freeze but this is limited when it comes to operating tape libraries like ejecting tapes and depositing new ones.

Our media db so far has 4550 tapes in its database. In addition this Netowrker server has a storage node.

We are still living with this problem and have a ticket open with Legato support # 3023296.

Any help is appreciated.

Thank you.

 

[TDun]

Try an mminfo -m, the snag is then finding a volume that is not in the volumes tab. If you do, then delete teh volume using
nsrmm -d -V volname

 

[jkalbus]

to TDun

we've got a similar problem - running networker 6.1.2 on Solaris with about 3600 tapes. I've checked the mminfo -m output and deletet two tapes with the nsrmm -d -V volname but it still is far too slow.

Any suggestions about "cleaning" the database? As I think this is the point where all the trouble comes together.

Johanes

 

[TDun]

Well, you could try an nsrim -X, t ocross check.
You could try compressing the media index (delete nsr\mm\cmprssd and then run nsrim)
You could try nsrck -m (my preference for a first step).

 

[LegLegato]

Well, TDun. I tried everything possible and nothing works. This is my latest try:

media database scavenging

- did a backup of bootstrap: savegrp -O -c <Networker server name> -G <group name>
and kept the bootstrap information handy
- shutdown Networker
- deleted the files under Program Files\nsr\tmp
- deleted ..\nsr\mm/.cmprssd and ..\nsr\mm\cmprssd if any
- deleted everything under ..\nsr\mm\mmvolume6
EXCEPT these files: VolHdr vol.0 ss.0 clients.0
In other words, do NOT delete these four files:
VolHdr vol.0 ss.0 clients.0
Deleted the rest under \nsr\mm\mmvolume6
- started Networker

No luck at all. Freezing is still there. I think this needs to be investigated from Legato engineering.

One more thing to add:
When this Networker Server was version 5.1 I didn't have this problem. We did an upgrade four months ago to new version 6.2 Not sure if this has to do with the conversion process of Media db.


Any thoughts?

 

[TDun]

They will probably frown heavily if you upgraded from 5.1 to 6.2 without going via 5.5.
Update this when you get an answer from Legato please. I haven't seen/suffered this, but wouldn't mind knowing the cause.

 

[wodstrup]

Hi,

We have the exact same problem with all our 4 True64 5.1 Legato 5.5 - 6.1 servers.

The amount of tapes is 10000

I have found a work-a-round:

In the nwadmin gui goto
options-customize-pollingintervals
and increase the media and index interval to maximum.

That way you only see the delay at startup of the gui

Yours
Frank Wodstrup
Sonofon
Denmark

 

[LegLegato]

Thanks wodstrup. Tried that. Didn't help at all. GUI freezes still the same. No answer so far. Legato support experienced the same thing. I ftp-ed my media DB to them and they build a Networker server with my media DB and they saw the same behaviour. They said that they are working on that. I still have no answer from them...

Looks like I will live with this for a while...


Any other idea is welcomed.

THanks to all of you.

 

[hkuhlmann]

Hi everybody,

we experienced exactly the same problem with nsr 6.2 on W2K Advanced Server (2 x 1 GHz Pentium III, 1GB RAM).

We now have 8.000 volumes in a virtual library tape system (FSC CentricStor) and will go to 15.000 in the next days.

When working with 8.000 tapes we had &quot;server freezing&quot;. We set the secondary polling interval to 300 and everything is fine again now. When upgrading to 15.000 we will have to set the value higher again. I will report wich value is okay.

Maximum should be 32.000 .

Hope this helps,
Heiko.

 

[LegLegato]

Wodstrup! I tried it today in another server and increasing the polling interval really helped. Thanks a million!!!!! I put 32000 and is working great. You are right, it delays only the first time but I can live with that.

Wow...

this is great.

THank you.