Errors from 4

[Inq]

Hi,

I am a newbie to Legato and had no joy with backing up without errors yet, so it's a plea to all you Legato experts. I will try and break my system down;

Legato Networker version 6.0.1 Build 174
Registered -
Autochanger/8 slots
ClientPak for Unix
Networker Module for Oracle UNIX client/1
Networker, Turbo/10
Update/7

Upon setting up the system I created a new group, new pool, new schedule and new label template for our company.

Legato Networker Server - 'Saturn' - NT 4.0 Server
File/Print Server - 'Triton' - NT 4.0 Server
File/Print Server - 'Apollo' - NT 4.0 Server
File/Print Server - 'Neptune' - NT 4.0 Server
Oracle DB Server - 'BamBam' - Solaris Unix
Oracle DB Server - 'Pebbles' - Solaris Unix

The hardware is a Quantum DLT7000 Tape Library with 7 70GB tapes and a barcode reader.

I installed the client software on the servers and did a name resolution to ensure they could see the server and the server could see them.

I put the clients into the new group/pool and did savesets = all. I added the clients into our new pool, new group, selected the label template and labelled all the tapes.

I set the browse/retention policy to a decade and selected NT compression and Unix compression directives - Everything else left as default.

I set the backup to run @ 3AM.

The group completed with errors;

Unsuccessful savesets

bambam:All 5 retries attempted
bambam:All nsrexec:authtype nsrexec
bambam:All savefs:unknown host name saturn
pebbles:All 5 retries attempted
pebbles:All nsrexec:authtype nsrexec
pebbles:All savefs:unknown host name saturn
pebbles:index 5 retries attempted
pebbles:index 02/10/03 08:31 nsrexec:Attempting to kill on remote save
pebbles:index aborted due to inactivity

The rest of the clients seem to backup fine. There are some errors;
unsyncronized client clock detected
NTUSER.DAT cannot open as in use
(Neptune) System State :\ aborted due to inactivity
(Neptune) Repair Disk :\aborted due to inactivity

The tapes were full and 100% full. It was also wanting to write the bootstrap and index to a different tape. I have browsed the forum and found a thread which hopefully will solve it.

So with the Unix machines not backing up at all and errors when backing up NT, I am unconfident of any kind of disaster recovery plan involving Legato at the moment. I realise it could be my lack of inexperience but I think the product could be a little more user friendly.

I realise that this new post is long, but I have tried to give as much information as possible to you guys to see if you can help me out. I have tried Legato.com but after I sent my daemon.log file 5 times, they don't seem interested or are too busy to help. I thought I would get some help from the true techies!

Regards,

Inq

 

[joeunixadmin]

Hi,

It looks like a DNS problem. NetWorker performs a reverse lookup on clients when attempting a backup. If you don't have DNS entries for the clients AND server you need to make them and include the reverse maps. If the entries are there, make sure that they are correct and that the reverse maps were created.

Check the DNS entries using nslookup.

Hope this helps.

-Joe

 

[kcongdon]

1. Try adding a HOSTS file or /etc/hosts file listing all the clients and the Legato server itself to the Legato server. Consider adding a HOSTS file or /etc/hosts file with an entry for the Legato server to every client. In theory a DNS server will handle name resolution properly in place of HOSTS files. In practice it doesn't always work right.

2. If the Legato server has more than one network interface, make sure all clients have all the interfaces listed in their /nsr/servers files.

Kate

 

[rgzimmer]

You do NOT want your browse policy set to a decade. You will run out of space on your server and/or have problems doing restores in about 6 months to a year.

Browse should be no more than 30 days unless you do a lot of restores, but even then making it longer is a gamble. Save set restores aren't that difficult.

As for the failures, look at the client set up and make sure that all aliases are listed on the client preferences tab. If one is missing and it is a valid host name on the client, you will get failures.

Are you using 6.01 on all machines? The "unsyncronized client clock detected" error is a message from 5.x I thought. I haven't seen it since we upgraded to 6.x.

And you shoudl seriously consider 6.1.3. It should be a free upgrade for you.

 

[Inq]

Thank you for all your posts.

Kate / Joe -
I have added as you said enteries in the the HOSTS and posted on all clients and the server. From the Unix machines it looked as though it was having problems with resolving the full DNS name it was trying to look for. Fingers crossed for the next backup - Thank you

rgzimmer -
Thank you for your advice on the browse period. I will take your advice and set it to 30 days (we don't do file restores that often). I checked the aliases and entered everything possible short of MAC address and inside leg measurement (Hopefully the host file will sort that one). I will check all client versions. That could be a problem as we had different versions. I will check to see if we can download the update from Legato as we don't have a support contract with them anymore. They did used to send us them at one point but stopped.

Does anyone else have a support contract with them, if so is it good value and are your response times quick?

Thank you again for your time in replying. I will know if it's looking healthier in the morning!

 

[wallace88]

For patches and updates, you can go to

http://www.legato.com/support,

then click on Updates and Patches.

 

[Inq]

Hi,

Looking at the backup this morning, there is a vast improvement in the Unix backup for computer 'Pebbles'. There are still some errors trying to backup the index. Everything else seemed to go fine, so will try and do a verify later to test it - Is this run like the NT client?

There is still a problem with 'bambam'.

The error;

Uncuccessful Savesets
bambam:/Attempting to kill on remote save
bambam:/saveinterrupted) exiting
bambam:/aborted due to inactivity

I am not sure but I think this could be due to the fact there are databases on here and they are interfering. Because we ran the backup on a full basis it ran into when the DB'd were started up. I will edit the time's and see if this cures it.

Wallace88 - Thank you for the advice. I will download the update today. Much appreciated.

We have setup incremental's all week and perform a full on Friday night. Would this be the best solution, or would you go for a level 9?

Thank you for all your help in getting this far.

Inq

 

[Inq]

Hi,

Looking at the backup this morning, there is a vast improvement in the Unix backup for computer 'Pebbles'. There are still some errors trying to backup the index. Everything else seemed to go fine, so will try and do a verify later to test it - Is this run like the NT client?

There is still a problem with 'bambam'.

The error;

Uncuccessful Savesets
bambam:/Attempting to kill on remote save
bambam:/saveinterrupted) exiting
bambam:/aborted due to inactivity

I am not sure but I think this could be due to the fact there are databases on here and they are interfering. Because we ran the backup on a full basis it ran into when the DB'd were started up. I will edit the time's and see if this cures it.

Wallace88 - Thank you for the advice. I will download the update today. Much appreciated.

We have setup incremental's all week and perform a full on Friday night. Would this be the best solution, or would you go for a level 9?

Thank you for all your help in getting this far.

 

[x2shy]

Personally, I wouldn't worry to much with the diffrent levels. Apart from the fact that they get way too confusing too quickly (like legato needs to have more confusing options in it!). It just gets messy.
Leave at full and incr, alot more logical and easier to maintain.
In regards to the issues on BamBam,
It may be an issue with parrellism on your groups.
Try putting BamBam in it's own groups one night,(or default and then see if it starts writing to tape)
Also look at the network connection to BamBam. Can you ping it? Can you nslookup? From both the backup server AND BamBam.
Give it a go.

 

[Inq]

Thank you x2shy for the advice - Agree that involving levels woould be more complicated than needs. Full and incr it is.

I have been running one group upto recently, but added another group for the bootstrap and index's (called bootstrap/Indexes). Even though all the clents are pointing to the same group (not default), they still want to write to the tape labelled default. I tried not labelling a tape as default but the backup says it's incomplete waiting for the tape to write to!!!

I will check the nslookup and the DNS connections between the servers. I am pretty sure these are ok though.

When you talk about parallelism in groups, the server is in the only one group I have in networker, apart from the bootstrap one. I will create another group and add it into there and try it.

Thanks again for your help.

Steven

 

[Inq]

Hi,

Well there has been a vast improvement. 'Hardly any errors at all'. I did get an error saying it couldn't write the index, but that was due to not having a tape labelled appropriately!

Would it be a good practice to have the bootstrap and index on the same tape as the rest of the backup?

It's looking alot better, having upgraded the Legato version and to WIndows 2000 Server, I was expecting some problems, but none appeared!

Thank's to all in the this forum.

Inq