• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 3323
  • Last Modified:

Windows - Delayed write failed

Hi,

We are getting this error intermittently on our file server, will be fine one day and everyone will be affected the next day. This was working fine before but this error has just started showing up. Happens to all Windows XP SP2 clients and the Server is running Windows 2003 Standard SP1. This tends to happen mostly with Excel 2003 files which are at least a couple of MB in size. Does not tend to affect any other applications. The file data is held on a SAN connected with fiber to the server. Things I have tried:

1. Changed physical port on switch for the server and also swapped the cable
2. Upgraded NIC drivers on server and client
3. Installed SMB hotfix from Microsoft
4. Disabled oplocks on client and machine
5. Upgraded firmware on the SAN. SAN monitor does not show any errors.
6. Connection on NIC of server changed to 100Mbps Full duplex and same setup on port on switch which connects to server.

We use Cisco 7941 phones for our users which acts as a switch and there is a connection to the desktop from the phone. The server and the clients are on different subnets. Clients tend to be connected through ethernet to the phone and have 100Mbps speed.

Error messages on client:
1. Windows was unable to save all the data for the file "network path". The data has been lost. This error may be caused by a failure of your computer hardware or network connection. Please try to save this file elsewhere.
2. In System Event Viewer on client, same error on server as well:
Event Type:      Warning
Event Source:      Ftdisk
Event Category:      Disk
Event ID:      57
Date:            6/27/2007
Time:            3:02:55 PM
User:            N/A
Computer:      MRP-GGATTO
Description:
The system failed to flush data to the transaction log. Corruption may occur.
3. Event Type:      Warning
Event Source:      MRxSmb
Event Category:      None
Event ID:      50
Date:            6/29/2007
Time:            3:34:05 PM
User:            N/A
Computer:      MRP-GGATTO
Description:
{Delayed Write Failed} Windows was unable to save all the data for the file \Device\LanmanRedirector. The data has been lost. This error may be caused by a failure of your computer hardware or network connection. Please try to save this file elsewhere.
4. On server:
Event Type:      Warning
Event Source:      MRxSmb
Event Category:      None
Event ID:      3019
Date:            7/23/2007
Time:            8:32:40 AM
User:            N/A
Computer:      FLPSRV01
Description:
The redirector failed to determine the connection type.
5. The File Server also acts as a Domain Controller and throws this MrxSMB error:
Event Type:      Error
Event Source:      MRxSmb
Event Category:      None
Event ID:      8003
Date:            9/6/2006
Time:            7:55:03 PM
User:            N/A
Computer:      FLPSRV01
Description:
The master browser has received a server announcement from the computer NS1 that believes that it is the master browser for the domain on transport NetBT_Tcpip_{D82A40DB-F4EE-4CB5-BE0B. The master browser is stopping or an election is being forced.

TIA,

Abhi
0
abhijitm00
Asked:
abhijitm00
  • 3
  • 2
1 Solution
 
BrughCommented:
If everything is running by NAME. IE DNS plays a big role in communication, then the Master Browser Error (MRXSMB) could be the cause.

Turn to the "computer browser" service off on NS1 and see if that helps out.

Also, check to see if the SAN device has "write caching" enabled. (doubt this could be it, but i woudl disable it if you are keeping yoru DB on it)

Do you have Dual NICs on the server and are both active and servicing Write Requests to the SAN?  If so, the Teaming may not be setup correctly, something to check.

0
 
abhijitm00Author Commented:
Hi Brugh,

Thanks for the input. We currently have write caching enabled on the SAN, I can try turning that off and see if that makes a difference. We don't have NICs teamed on the server. Is there a way to check to see if its a network/subnet problem? Thanks.
0
 
BrughCommented:
As long as routing works across the subnet, ie, you can access a node on the other subnet, you should be fine.

I would do:
 - a ping/tracert from a client machine to Application Server
 - a ping/tracert from the Client Machine to the SAN
 - a ping/tracert from the Application Sever to the SAN.

I would do this periodically while the issues are occuing.  Hopefully you will be able to see if network problems are in play while the write-caching is turned off.

Another "cleanup" item that you can do is to upgrade to Server 2003 Sp2. (just make sure you check with any application vendors to make sure its an approved upgrade)
0
 
BrughCommented:
"problems are in play while the write-caching is turned off."

Should read, "in play while the "delayed write" failures are occuring."  :)
0
 
abhijitm00Author Commented:
No lost packets when opening/copying files from the server on either the client, server or SAN.
0

Featured Post

The 14th Annual Expert Award Winners

The results are in! Meet the top members of our 2017 Expert Awards. Congratulations to all who qualified!

  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now