Video Screencast Help
Search Video Help Close Back
to help
New in the Rewards Catalog: Vouchers for "Symantec Technical Specialist" and "Symantec Certified Specialist" exams.

network connection timed out(41)

Updated: 21 May 2010 | 23 comments
ruffy's picture
0 0 Votes
Login to vote

hi

i backup the mailboxes on an exchange server. all mailboxes aborts with network connection timed out (41).
i diveded the mailboxes in two sections (a-c, and d-g) and four attempts to backup. timeouts are at 7200.
this is a log of a backup policy:

17.10.2007 08:24:26 - end writing; write time: 00:29:25
network connection timed out(41)
17.10.2007 08:24:27 - Error bpsched(pid=5760) suspending further backup attempts for client dsexc120, policy DSEXC120_MAILBOXES_A, schedule daily because it has exceeded the configured number of tries
17.10.2007 08:24:28 - Error bpsched(pid=5760) backup of client dsexc120 exited with status 41 (network connection timed out)

the exc-srv has a direct network connection to the backup server (switch is between the both).
nic driver are actual on both too.

whats the problem?

discussion Filed Under:

Comments

Casey King's picture
17
Oct
2007
0 Votes 0
Login to vote

I am not exactly sure what your environment looks like but these are some things to consider when doing backups
 
Are your exchange servers on the same network segment?
 
If they are on the same network segment then there is definitely going to be some contention for the bandwidth. 
 
What other backups are running at the time you are trying to do your exchange backups?
 
I had a similiar problem with my SAP backups that were giving me 41 errors.  I had to review what I was backing up, the times of the backup, and where these servers were on the network.  I was able to break down my backups by average run times, where they were on the network segments, and separated them into different policies with runtimes that overlapped by about 30 minutes.  I also moved other nightly backups to a later time since I moved my SAP backups to an earlier time.  Those backups are still completing without errors.
 
 
 
I know exchange backups are very important to the company, but I would try to look at the environment as a whole.  What is taking place at the same time as your exchange backups....if it is permitted you might disable the policies that run at the same time to see if you get better results.
 
 
 
ruffy's picture
17
Oct
2007
0 Votes 0
Login to vote

hi casey king:

my exc and backup srv are on the same network segment (both have a nic with the same segment). their realy network segments are defferent (exc 192.168.160.xyz, backup network 192.168.250.xyz and backup srv 10.0.0.xyz)

i runs the backup policy for couple minutes today!!! it runs no other policies at the same time.

any other ideas?

Casey King's picture
17
Oct
2007
0 Votes 0
Login to vote

Let me see if I am on the same page as you:
 
1. Exchange and Master/media servers are on the same physical network segment
2. Exchange and Backup servers are on a different addressing schemes
 
I guess my question at this point is
 
1. Has this backup ever worked
2. Are you able to ping your exchange servers from the master server
3. Do you have host files set (not that I personally recommend this...although NB support does)
 
Also, let us know your environment
1. Platforms
2. Version of NB
3. Any other important information
ruffy's picture
17
Oct
2007
0 Votes 0
Login to vote

no, thats wrong:

backup server has two nic (own and for the backup). the same has exc (own and for the backup).


1. yes thes works without errors for a couple of months
2. im able to ping the exc (with ip and name).
3. yes im working with hosts

Plattform: windows 2003 server enterprise sp2
Version of NB: 5.0 MP7
other: do not know which

Casey King's picture
17
Oct
2007
0 Votes 0
Login to vote

Sounds like everything is right....
 
I am curious if you are able to access the client properties through the admin GUI.  Have you checked the clients themselves to see if the client tracker believes it is still running a backup?  Try to stop/start the client services on the exchange server....beyond that I am not sure what else to check
ruffy's picture
17
Oct
2007
0 Votes 0
Login to vote

thats my problem!!!

i stoped the client already without any changes.

i will install mp7 on the client and try again!

MattS's picture
17
Oct
2007
0 Votes 0
Login to vote

Can you do this from the client to the server and from the server to the client?
 
telnet *client or server name* bpcd

ruffy's picture
17
Oct
2007
0 Votes 0
Login to vote

it works.

MattS's picture
17
Oct
2007
0 Votes 0
Login to vote

Well thats good.
 
Have you tried looking through this guide?

ruffy's picture
18
Oct
2007
0 Votes 0
Login to vote

hi MattS

thanks for this document. its realy good. i will proof it all and write the results and mayby the resolve of my problem

sdw303's picture
22
Oct
2007
0 Votes 0
Login to vote

Ruffy, you say the timeout is set to 7200 seconds.  Is this on the Exchange client only, or on the media server only, or on both?
 
I believe that for MS Exchange policies (in fact any "agents", e.g. DB2, Domino, Oracle, SQL, etc), any changes to the Client Read Timeout should be set on both the client *and* on the media server performing the backup to tape/disk.  This is because "Client Read Timeout" is the amount of time that a "Media Server" and the "NetBackup Agent" (in this case Exchange) will allow for a backup to not respond, i.e. it is the amount of time that may pass before the "media server" OR the "client agent" gives up and decides that either the client or the backup in general has failed.
ruffy's picture
23
Oct
2007
0 Votes 0
Login to vote

hi sdw303

you are right, the "client read timeout" should changes only on one (master server or client).
but "file browse timeout" can be changed only on the master server.

yes, the timeouts are on both the same. i read the help of the general tab on the client in the restore client and increase the client read timeout and file browse timeout to the maximum (32400). i will test it till tomorrow.

thanks for the time being.

sdw303's picture
23
Oct
2007
0 Votes 0
Login to vote

Hi Ruffy,
The "browse" timeout only controls how long the client side restore window will wait when browsing the catalog for restores.  This "browse" timeout has no baring on backups.  Normally 300 seconds is enough for most restore sessions, but if th ecatalog is huge, or the client has many recent backups then the browse timeout may need to be extended.  I've never had to though.
HTH,
Dave.
ruffy's picture
23
Oct
2007
0 Votes 0
Login to vote

hi dave

thanks for this info. i will change the browse timeout back to 300. client timeout i will test anyway. i became status 41 only at the mailboxes backup.

sdw303's picture
23
Oct
2007
0 Votes 0
Login to vote

Have you definitely set the CLIENT_READ_TIMEOUT on both the MS Exchange NetBackup client and also on the NetBackup media server that is performing the backup to tape?
ruffy's picture
23
Oct
2007
0 Votes 0
Login to vote

oh **bleep**,

you are right again. the option i thout is client read timeout is list file timeout on the exc client. where can i do this on the client?

sdw303's picture
23
Oct
2007
0 Votes 0
Login to vote

I'm using v5.1 MP6 Windows, but it should be similar in v6.x...
 
For the client that is the MS Exchange server:
Host Properties - Clients - select client name, right click, select "Properties" - then look in the "Timeouts" tab, and the first option (for me) is "Client read timeout".
 
For the media server:
Host Properties - Media Server - select server name, right click, select "Properties" - then look in the "Timeouts" tab, and the "Client read timeout" (for me) is the top right option.
 
HTH.
 
PS. I have set all five of our media servers to 3600, and both nodes of our MS Exchange server cluster to 3600 seconds too.  We only allow mailboxes to grow to 40 MB mostly, but a few are larger, and it's these few larger ones that still timeout occaisionally whilst the server is busy with all the mailbox backups at the weekend, but when they retry later when it's quiet they work ok with a setting of 3600 seconds.  IMHO, if your mailboxes are typically larger, they you may need to increase from 3600 to 5400 or even 7200 - but I would have thought that you'll get fewer and fewer status 41 mailbox backup with each increase.  In the end you may even have one or two mailboxes that still refuse to work, in which case I would suggest that the contents of the mailboxes be pruned to sensible levels.
ruffy's picture
23
Oct
2007
0 Votes 0
Login to vote

hey thanks sdw303,

i did not know about this!!!

i changed it from 300 to 7200, because our customer (exc from one customer) and i do not have permissions to change the size or to reduce it. i will get you the result.  thanks a lot!

ruffy's picture
24
Oct
2007
0 Votes 0
Login to vote

it works!!!

thanks. all infos from you, sdw303, was helpfull. thanks a lot.

it can be closed.

Stumpr's picture
24
Oct
2007
0 Votes 0
Login to vote

Another thing to watch is fragmentation of the disk drives. Run defrag once in a while on window clients.
 
 
 
 

Bob Stump VERITAS - "Ain't it the truth?" Incorrigible punster -- Do not incorrige

ruffy's picture
26
Oct
2007
0 Votes 0
Login to vote

thanks for help stumpr, but it does work with the high timeout of 7200.



Message Edited by ruffy on 10-26-2007 12:01 PM

Stumpr's picture
26
Oct
2007
0 Votes 0
Login to vote

yes, I'm glad to hear that. But remember that in future it is best practice to check fragmentation of client's disk drives.
 
 

Bob Stump VERITAS - "Ain't it the truth?" Incorrigible punster -- Do not incorrige

ruffy's picture
26
Oct
2007
0 Votes 0
Login to vote

thanks, i will do it.