Backup Of H:\ Drive failing with status 42(network read failed) for windows 2000 client

[Venu_S1]

Hi All,

Backup Of H:\ Drive failing with status 42(network read failed) for windows 2000 client.

The rest of the streams completing successfully.

nbu version 6.5.6

please chelp..

[RamNagalla]

hi,

how much of data that you are trying to backup with H drive?

does it writing any data before failing or just failing with out writing any data?

did you try increasing client read time out and client connect timeout values in media server?

please post the detail status of the failed job

[Venu_S1]

Hi Nagalla

The data size of H:\ drive is 269 GB

its getting failed after writing around 80 GB of data.

Apr 5, 2013 8:04:13 AM - estimated 0 kbytes needed
Apr 5, 2013 8:04:15 AM - started process bpbrm (pid=10064)
Apr 5, 2013 8:04:22 AM - connecting
Apr 5, 2013 8:04:23 AM - connected; connect time: 0:00:00
Apr 5, 2013 8:04:34 AM - begin writing
Apr 5, 2013 6:37:07 PM - Error bpbrm (pid=10064) db_FLISTsend failed: network read failed (42)
Apr 5, 2013 6:38:43 PM - end writing; write time: 10:34:09
network read failed  (42)
 

not changed any values on media server..

[RamNagalla]

pleae go ahead and increase the client read timeout vault to 1800 or more than that... and try backup again.. 

[mph999]

I would argue that 1800 for a client read timeout is too high.

You need to fix the problem, not work around it.

Timeouts 'can' be a fix, for example some DB backups require 'large' timeouts, but for general clients (which I presume this is) a timeout adjustment may not be the best fix.

In this case, the issue is a communications issue between the  media server and the master.

bpbrm tried to send catalog information to bpdbm but failed.  Not sure a client read timeout would make any difference.

I'd start this issue by getting your network team to take a look, as 99% of the time, that is where the problem is most likely to be.

Big question, what changed, did this ever work, any recent patches etc ...

Martin

[Venu_S1]

Please find the attached bpbkar log with verbose 5. might be you will get the idea about the error..

[mph999]

http://www.symantec.com/docs/TECH6127

I suspect the main issue is started between the media server and the master, and the client errors are because of this.

You would need to find the time of the first error, does the media server show an error between it and the master before or after the errors shown by the client.

Looking at what you posted, the bpbrm process  shows an error at 06.37pm and the bpbkar log at 09.04.  I don;t know if there is a time difference between the servers, this would need to be taken into account if there is.

So , based on what I know, I would say the issue is a network problem between the master and the media.

[Possible]

Venu,

<16> dtcp_write: TCP - failure: send socket (736) (TCP 10054: Connection reset by peer)
 <16> dtcp_write: TCP - failure: attempted to send 157 bytes
 <16> dtcp_write: TCP - failure: send socket (736) (TCP 10054: Connection reset by peer)

First suggestion is to have network team involved here.

You can go one by one as below.

On windows systems, the client bpbkar log may contain a 10054
"Connetion Reset Error" error (Usually indicates hardware error)
somewhere between the Netbackup Client and server,the connetcion was
reset. When Netbackup receives this error, it can not continue the backup.
This error can be contributed by following :
1)Hickup in Network
2)A bad NIC on client
3)A bad NIC on NBU server
4)Faulty routers
5)Other application that interfere NBU connections.

Thanks.

[Possible]

pls update ur status once as suggested...

[Mark_Solutions · Accepted solution]

This backup obviously takes a very long time which can cause a lot of issue, timeouts, keep alive etc. etc.

If a firewall is between the two that can also cause the same type of issue as they also tend to have the 2 hour keep alive timeout.

As a test you could enable checkpoints for the backup and set it to 30 minutes

This forces communication between master / media / client every 30 minutes which may overcome things for you - if this works either stick with it or try and identify where the keep alive setting comes from.

Hope this helps