nwrecover or recover

[mouse123]

Hi All

I am not able to start either nwrecover or recover in unix. but I can run nwadmin and nwbackup without anyproblem

#nwrecover -s system(servername)
#recover -s system(servername)

What could be the wrong?

Thanks in Advance

 

[mouse123]

I did restore 2 weeks old full backup.

 

[TDun]

To be honest, it sounds more like it is the client file index for production1.
Try running nsrck -F and then nsrim -X.
Then try the recover, specifically using recover, not nwrecover.

 

[mouse123]

I did the same, but still i am getting the same message

./recover -s production1
recover: Cannot start session with server production1: Unable to receive

Now I don't what to do? My data is so critical. Anyother help please

Thanks in advance

 

[RMinNJ]

I know your on the server itself and you tried it but if its
not an index problem it still sounds like a name resolution
problem. Beside's using the fully qualified domain name, I'd
make sure the reverse lookup of the server's ip resolves to
the short or fully qualified name ie:

nslookup x.x.x.x

name: production1.xyz.com

 

[mouse123]

Hi Rmaiello

all the places I am using name like production1.xyz.com

Name resolution is not a problem here

thanks

 

[605]

Puh,

you have a real strange problem. Actually, you can even recover without a client file index. Please choose an appropriate save set an then start

# recover -s server_name -S ssid

Let us see, if this is possible. If so, it again points to
a problem with the client's file index (or still name resolution).

Additionally you could run these commands:
nsrls
nsrinfo clientname |more
just to see whether you can read the CFI

Regards

 

[mouse123]

Hi 605,
when i run the following command

production1:/oracle/ora81/stage_tmp/lsm/aix41#./nsrls

/nsr/index/production1.xyz.net/db: 10665 records requiring 2.4 MB
/nsr/index/oraprod.xyz.net/db is currently 100% utilized


Is it something wrong in the above result ?

Ashok

 

[605]

As far as i know, the LSM product is just capable to backup the local files - no other clients. Why do you see 2 client
file indexes then? - one with no obvious contents, except that it is to 100% used.

Did you rename the machine?

 

[mouse123]

605,

I did rename the server a month ago, after that also recover & nwrecover was working fine.

Now I have removed the other name in the server. here is the new result

/nsr/index/production1.xyz.net/db: 11207 records requiring 2.5 MB
/nsr/index/production1.xyz.net/db is currently 100% utilized

thanks

 

[605]

What you should do is make sure that all names are listed in the client's alias list. This should 'link' all the client file index databases.

However, this still points to a name resolution problem.
Make sure that the name resolution will be unique.

 

[mouse123]

605,
In the clients options, I did specified in

Remote access:
production1.xyz.net
production1
Aliases:
production1.xyz.net
production1

Is there anyother place i need to link the clients

Again

When I nslookup on these servers names

#nslookup
Default Server: ntprimary.xyz.net
Address: 10.10.x.x
> production1
Server: ntprimary.xyz.net
Address: 10.10.x.x
Name: production1.xyz.net
Address: 10.10.x.102

> production1.xyz.net
Server: ntprimary.xyz.net
Address: 10.10.x.x

Name: production1.xyz.net
Address: 10.10.x.102

Thanks in Advance

 

[605]

So far so good. Now you need to make sure that reverse lookups are also o.k.

 

[mouse123]

production1:/#nslookup
Default Server: ntprimary.xyz.net
Address: 10.10.x.x

> 10.10.x.102
Server: ntprimary.xyz.net
Address: 10.10.x.x

Name: production1.xyz.net
Address: 10.10.x.102
>exit

 

[605]

This seems to be o.k. but there could still be problems due to a lot of other issues. This is not easy to investigate.
So let us forget about the client file index for a while.

Did you ever try a save set recovery what i mentioned two days ago?

 

[mouse123]

I try to recover from my old full backup, but


Hi 605,

I did tried the following

-run mminfo -av
-stop all the daemons
-copy the nsr directory different filesystem
-deleted the /nsr/index/client_name
-start NW

It does created an empty client file index database

then I tried

./recover -s production -S 2427

the following message

recover:cannot start session with server production1: unable to receive

Any other tips please

Thanks in Advance

 

[605]

Sorry, but you confuse me ... and yourself.
Now you are dealing even with a new name: "production" as
NW server.

Additionally, you do unnecessary steps - why copying NW ?
You just wanted to recover something. This does not change
a database at all.

Why did you delete the client file index ? - By specifying the save set ID (ssid) you simply bypass the client file index.

Of course, as NW knows about the client from the resource files, he will re-create the (empty) CFI if the NW daemons
will be restarted.

Friendly, i think you first need some explaination about how NW databases work in general. Let me suggest that you
download the following document:

ftp.legato.com/pub/NetWorker/German/tdocs/tids_2002.pdf

and start to read it.

------

BTW if the save set you want to recover is a oracle save set, i cannot be of so much help. However, your error message could also mean that the Oracle recovery catalog
requesting a save set that is still available in the recovery catalog but does not exist in the media index any
longer.

 

[mouse123]

Sorry about my previouse posting, My client name & server name is production1.xyz.net

Thanks for your documents

whatever you posted as below

In this case please vary the procedure as follows:
- run mminfo -av -r "name=index"
- stop all daemons
- copy the nsr directory
- delete the /nsr/index/client_name
- start NW.
It will recreate an empty client file index database
- run "recover -S ssid#"
Where ssid# is the number of the last full index backup

which I have followed.

even to restore the save set, I am not able to execute recover

this is what i have done for recover save set

./recover -s production1.xyz.net -S 2427

the following message

recover:cannot start session with server production1: unable to receive

 

[jkalbus]

Just maybe one point - are you sure your routing is alright?
Do you have more than one network on your backup server and / or backup client?
If yes, maybe you should try the name of the machine in the different network.

Johanes

 

[mouse123]

Johanes,

It was working without any problem,

we didn't change any network, still it is not working

Thanks