You cannot access the existing File Share resources on a Windows Server 2003 failover cluster

http://support.microsoft.com/kb/961657

 解决Operating system error 64(The specified network name is no longer available.的错误，设置SessTimeout

On a Windows Server 2003 failover cluster, you may experience one or more of the following symptoms.

Note The "Workaround" section also applies to non-cluster scenarios where you cannot access a file share.

Symptom 1

You cannot access the existing File Share resources even if these resources are online according to the Cluster Administrator snap-in. Additionally, the following error may be logged in the event log:

Date: date
Time: time
Source: ClusSvc
Type: Error
Event ID: 1055
User: N/A
Computer: ServerName
Description:
Cluster File Share resource ' ShareName' has failed a status check. The error code is 64.

Symptom 2

One of the following error messages is logged in the Cluster.log file.

Error message 1

00000930.00000af8::{2003/01/01 23:00:00.001} File Share ShareName: Share has gone offline, Error=64!
00000930.00000934::{2003/01/01 23:00:00.001} File Share ShareName: Share has gone offline, error=64!

Click on Start, Run, cmd.exe
Type "Net HelpMsg 64"
"The specified network name is no longer available."

Error message 2

Description: Cluster File Share resource ' ShareName' has failed a status check. The error code is 32.

Click on Start, Run, cmd.exe
Type "Net HelpMsg 32"
"The process cannot access the file because it is being used by another process."

Symptom 3

You cannot access local shares.

Back to the top | Give Feedback

Cause

When the Cluster service (ClusSvc.exe) cannot perform an IsAlive test against the File Share resource, event ID 1055 is logged. Usually, the IsAlive test fails because the Cluster service cannot connect to the Server service (Srv.sys). This failure causes the Server service to be unable to keep up with the demand for the network work items that the network layer of the I/O stream queues.

For more information about the IsAlive test, click the following article number to view the article in the Microsoft Knowledge Base:

914458

(http://support.microsoft.com/kb/914458/ )

Behavior of the LooksAlive and IsAlive functions for the resources that are included in the Windows Server Clustering component of Windows Server 2003

You may experience these symptoms for one or more of the following reasons:

• A slow disk I/O stream causes disk response time to increase. (This is the most common cause of the symptoms.)
• The server is under high load.
• An application or a service causes a deadlock in the Server service.
• Device drivers on the server are outdated. For example, the network adapter driver is outdated.
• Applications that use file system filter drivers are running on the server. Typically, the following applications use file system filter drivers:
  • Antivirus software
  • Backup applications
  • Quota management applications
• A tape device on a server is attached to the same storage area network (SAN) that the cluster is using. The tape device driver decreases the disk queue length on the SAN. This action affects each server that is attached to the SAN. You can increase the disk queue length to avoid this situation.

Back to the top | Give Feedback

Workaround

Important This section, method, or task contains steps that tell you how to modify the registry. However, serious problems might occur if you modify the registry incorrectly. Therefore, make sure that you follow these steps carefully. For added protection, back up the registry before you modify it. Then, you can restore the registry if a problem occurs. For more information about how to back up and restore the registry, click the following article number to view the article in the Microsoft Knowledge Base:

322756

(http://support.microsoft.com/kb/322756/ )

How to back up and restore the registry in Windows

Note The "Workaround" section also applies to non-cluster scenarios where you cannot access a file share.

To work around this issue, follow these steps.

Step 1: Tune the Server service to increase capacity

Note Follow this step if the server has more than 1.5 gigabytes (GB) of system memory. Otherwise, go to the "Step 2: Tune the Workstation service to increase capacity" section.

To follow this step, use one of the following methods.

Method 1

1. Click Start, click Run, typeregedit.exe, and then click OK.
2. In Registry Editor, locate and then right-click the following registry subkey:
   HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\lanmanserver\parameters
3. Create a MaxFreeConnections registry entry, and then set the value to 1000 (hexadecimal). To do this, follow these steps:
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MaxFreeConnections, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type1000 under Value data, and then clickOK.
4. Create a MinFreeConnections registry entry, and then set the value to 64 (hexadecimal). To do this, follow these steps:
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MinFreeConnections, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type64 under Value data, and then clickOK.
5. Create a MaxRawWorkItems registry entry, and then set the value to 200 (hexadecimal).
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MaxRawWorkItems, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type200 under Value data, and then clickOK.
6. Create a MaxWorkItems registry entry, and then set the value to 2000 (hexadecimal).
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MaxWorkItems, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type2000 under Value data, and then clickOK.
7. Create a MaxMpxCt registry entry, and then set the value to 800 (hexadecimal).
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MaxMpxCt, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type800 under Value data, and then clickOK.

Method 2

1. Click Start, click Run, typeNotepad, and then click OK.
2. Copy the following text, and then paste it into Notepad:

   Windows Registry Editor Version 5.00 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\lanmanserver\parameters]
   "MaxFreeConnections"=dword:00001000
   "MinFreeConnections"=dword:00000064
   "MaxRawWorkItems"=dword:00000200
   "MaxWorkItems"=dword:00002000
   "MaxMpxCt"=dword:00000800

3. Save the file as .reg file type, and then exit Notepad.
4. In Windows Explorer, double-click the file to import these registry settings.

Step 2: Tune the Workstation service to increase capacity

To follow this step, use one of the following methods.

Method 1

1. Click Start, click Run, typeregedit.exe, and then click OK.
2. In Registry Editor, locate and then right-click the following registry subkey:
   HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\lanmanworkstation\parameters
3. Create a MaxCmds registry entry, and then set the value to 1000 (hexadecimal).
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type MaxCmds, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type1000 under Value data, and then clickOK.

Method 2

1. Click Start, click Run, typenotepad, and then click OK.
2. Copy the following text, and then paste it into Notepad:

   Windows Registry Editor Version 5.00 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\lanmanworkstation\parameters]
   "MaxCmds"=dword:00001000

3. Save the file as .reg file type, and then exit Notepad.
4. In Windows Explorer, double-click the file to import these registry settings.

Step 3: Enable the SessTimeout registry entry setting for the Workstation service

Note Follow this step only if you have Distributed File System (DFS) shares.

To follow this step, use one of the following methods.

Method 1

1. Click Start, click Run, typeregedit.exe, and then click OK.
2. In Registry Editor, locate and then right-click the following registry subkey:
   HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\LanmanWorkstation\Parameters
3. Create a SessTimeout registry entry, and then set the value to 300 (Decimal).
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type SessTimeout, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickDecimal under Base, type300 under Value data, and then clickOK.

Method 2

1. Click Start, click Run, typenotepad, and then click OK.
2. Copy the following text, and then paste it into Notepad:

   Windows Registry Editor Version 5.00 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\lanmanworkstation\parameters]
   "SessTimeout"=dword:0000012c

3. Save the file as .reg file type, and then exit Notepad.
4. In Windows Explorer, double-click the file to import these registry settings.

For more information about the SessTimeout registry entry, click the following article number to view the article in the Microsoft Knowledge Base:

102067

(http://support.microsoft.com/kb/102067/ )

SESSTIMEOUT Information

Step 4: Improve disk I/O performance

Disk fragmentation causes disks to perform additional data-seeking operations and slows down the rate of data transfer. To improve disk I/O performance, follow these steps:

1. Perform a defragmentation analysis on high-load volumes.
   
   Disk fragmentation can cause slow disk I/O. If the disk cannot sustain the network load, the response time of the disk increases. You can perform a defragmentation analysis on high-load volumes to determine whether the volumes should be defragmented.
   
   To perform a defragmentation analysis, follow these steps.
   
   Important Do not perform a defragmentation analysis on a volume that has active user operations.
   1. Click Start, clickRun, type Dfrg.msc, and then clickOK.
   2. In the Disk Defragmenter snap-in, click the volume that you want to analyze, and then clickAnalyze.
2. If you receive the following message after the analysis is complete, you should schedule a defragmentation of the volume or rebuild the volume.
   Analysis is complete for: ( VolumeNumber:)
   You should defragment this volume.
   If you have 200GB or more of data on the volume, it is faster to rebuild the volume than to defragment the volume. To rebuild the volume, follow these steps:
   1. Create a new logical unit number (LUN).
   2. Copy the data on the volume to the LUN.
   3. Copy the data back to the volume from the LUN.
   After you rebuild the volume, you should run a third-party disk defragmenting utility to make sure that the volume is not highly fragmented.
3. Run the Chkdsk.exe utility against the volumes.
   
   Note Follow this step only if the defragmentation analysis finishes quickly and without any error.
   
   You cannot schedule a disk check task on shared disks when the system starts. Therefore, interactively running the Chkdsk.exe utility is the only method that is supported on a Windows Server 2003 failover cluster.
   
   Be aware that if severe corruption exists on the volume, running the Chkdsk.exe utility can take the disk offline and disconnect all users. When you interactively run the Chkdsk.exe utility, you can stop the utility if the Chkdsk.exe utility produces severe errors. If the Chkdsk.exe utility exits, you must rebuild the volume. Disk corruption can cause an I/O bottleneck when read or write operations occur on the volume. Remember that if you run the Chkdsk.exe utility against a large volume, it may take a very long time for the Chkdsk.exe utility to finish. The actual time can vary from one day to two to three weeks. Therefore, before you run the Chkdsk.exe utility, you should consider scheduling downtime for the volume.
   
   To run the Chkdsk.exe utility against an online volume, type the following command at a command prompt, and then press ENTER:
   Chkdsk.exe X:
   Note The placeholder X represents the drive letter. If you have multiple volumes to check, you must run a separate command for each volume.

Back to the top | Give Feedback

Step 5: Apply the latest hotfix for certain Windows components

Apply the latest hotfix for the following Windows components:

• Server service driver
• Windows Kernel
• SMB mini-redirector driver
• TCP/IP driver

To obtain the latest hotfixes for these components, follow these steps:

1. Visit the Microsoft Support Center. To do this, visit the following Microsoft Web site:
   http://support.microsoft.com/

   (http://support.microsoft.com/)
2. Use the "Srv.sys kbqfe" string together with the "Windows Server 2003" string to search for articles about the Server service driver. For example, Microsoft Knowledge Base article 950298 appears in the search results.
3. Use the "NTOSKrnl.exe kbqfe" string together with the "Windows Server 2003" string to search for articles about the Windows kernel driver. For example, Microsoft Knowledge Base article 942835 appears in the search results.
4. Use the "MrxSmb.sys kbqfe" string together with the "Windows Server 2003" string to search for articles about the SMB mini-redirector driver. For example, Microsoft Knowledge Base article 925903 appears in the search results.
5. Use the "TCPIP.sys kbqfe" string together with the "Windows Server 2003" string to search for articles about the TCP/IP driver. For example, Microsoft Knowledge Base article 950224 appears in the search results.

Step 6: Obtain the latest device drivers

Outdated device drivers can reduce server performance. Therefore, we recommend that you follow these steps:

1. Obtain the latest device driver updates. For example, obtain the driver update for the HP Solid State Drive (SSD). You can use the IBM UpdateXpress tool to update device firmware. Frequently, manufacturers release updated device drivers to resolve bottleneck issues. Contact hardware vendors to obtain the latest driver updates and to verify that you are using the current version of the device firmware. It is especially important to do this for disk subsystem hardware.
2. If you use SANs, follow these steps:
   1. Check the driver configuration. For example, check the driver and firmware for SANs and for host bus adapters (HBAs).
   2. Verify that you have the most recent version of the Storport storage driver (Storport.sys).
   3. Check the following items:
      • HBA queue depth setting
      • Multipathing software
      • Fibre Channel connection

Step 7: Disable file system filter drivers if it is possible

Usually, quota management applications, open file agents, and file replication applications use file system filter drivers. Therefore, if you disable these applications, you disable the file system filter drivers. After you take this action, you can determine whether the problem is resolved.

You can also configure the antivirus software on the server to disable real-time scanning of all files. If you cannot disable real-time scanning of all files, we recommend that you configure the antivirus software so that it scans only incoming files and does not scan any of the following files:

• Page files
• .vhd files
• .tmp files
• .shd files
• .spl files

Step 8: Understand the load on the server, and collect system information in case this problem occurs again

After you perform the previous seven steps, the problem should be resolved. However, you should make sure that you have an idea of the load on the server after the server is back in production. Additionally, you should make sure that you collect system information in case this problem occurs again.

You can collect general performance data from the server to obtain an idea of the load on the server.

Note We recommend that you collect the data for at least three hours per day.

To collect general performance data from the server, follow these steps.

1. Obtain an output from the Net Files command and from the Net Sessions command. To do this, one at a time, type the following commands at a command prompt, and then press ENTER.
   
   Note In the following commands, the placeholder ServerName represents the name of the server on which this problem occurs. The placeholderMMDDYEAR represents the date in the MMDDYear format. For example, you type01012008 for MMDDYEAR.
   Net.exe Files > C:\NetFiles_ServerName_MMDDYEAR.log
   
   Net.exe Sessions > C:\NetSessions_ServerName_MMDDYEAR.log
2. Collect the Performance Monitor log.
   
   Note Follow the appropriate steps, depending on the version of Windows Server 2003 that is running on the computer.
   
   Windows Server 2003, x86-based versions
   1. Obtain the PerfWiz.exe utility. To do this, visit the following Microsoft Download Web site:
      http://download.microsoft.com/download/b/0/e/b0e8dae9-b8ba-49a8-8ae7-5a967a7421a1/PerfWiz.zip

      (http://download.microsoft.com/download/b/0/e/b0e8dae9-b8ba-49a8-8ae7-5a967a7421a1/PerfWiz.zip)
   2. From a remote computer, run the Performance Monitor Wizard utility (PerfWiz.exe), and then clickNext.
   3. Next to Monitoring Computer, type the name of the local computer, and then clickNext.
   4. Click Create New Log, and then clickNext.
   5. Click Standard Perfmon, and then clickNext.
   6. Type the name of the server that encounters the problem, and then clickNext.
   7. Next to Log Name, type the name that you want to use for the log, and then clickNext. You can use the default size of 200 MB that appears next toLog file size for the log size. If you specify a different size, make sure that you do not specify a size that is larger than 250 MB. If the log size is larger than 250 MB, it is almost impossible for the system to read the log because of the delay that occurs when the system is loading the performance counters.
   8. Take one of the following actions:
      • Under Average Time to issue, type 6 hours.
      • Under Sample Interval, type 300 seconds.
   9. Click Next.
   10. Click START, and then clickNext.
   11. Click Finish.
   12. Run the PerfWiz.exe utility to create another Performance Monitor log. To do this, use the standard log profile, and use the following settings:
       • Set Log file size to 150 MB.
       • Set Average Time to issue to 1 hour.
   Windows Server 2003, x64-based and Itanium-based versions
   1. From a remote computer, click Start, click Run, type PerfMon, and then click OK.
   2. In the Performance snap-in, expand Performance Logs and Alerts.
   3. Right-click Counter Logs, and then clickNew Log Settings.
   4. In the New Log Settings dialog box, type a name for the new log, and then clickOK.
   5. In the LogName dialog box, clickAdd Counters.
      
      Note The placeholder LogName represents the name of the log that you created in the previous step.
   6. In the Add Counters dialog box, clickSelect counters from computer, and then specify the name of the server that encounters the problem. Specify the server name in the following form:
      \\ ServerName
   7. Under Performance object, selectProcess.
   8. Click All counters, clickAll instances, and then click Add.
   9. Repeat the previous two steps to add the following objects:
      • Cache
      • Memory
      • Objects
      • Paging File
      • LogicalDisk
      • NBT Connection
      • Network Interface
      • PhysicalDisk
      • Processor
      • Redirector
      • Server
      • Server Work Queues
      • System
      • Thread
      Note For some objects, the All instances option does not apply.
   10. In the Add Counters dialog box, clickClose.
   11. In the LogName dialog box, type600 next to Interval, and then selectseconds next to Units.
   12. Click the Log Files tab, selectBinary Circular File under Log file type, and then click Configure.
   13. In the Configure Log Files dialog box, type250 under Limit of, and then clickOK
       
       Note If you set the size limit to 250 MB, you can see enough history to recognize a trend. Notice that the log can become very large. However, the log will be automatically compressed to about 20 percent of its original size. Therefore, if the log reaches the maximum size, the log will be automatically compressed to around 50 MB.
   14. In the LogName dialog box, clickOK.
   15. Repeat steps c through o of this subsection to create another log. However, type5 next to Interval in step k.
3. Obtain the description of the files that are generally opened on the server and the description of the roles of the server. For example, a server may act as a file server or as a terminal server.

You can use the Network Monitor utility to capture network traffic in case this problem occurs again.

Note Capturing network traffic is also known as capturing the network sniffer trace.

To download the Network Monitor utility, visit the following Microsoft Download Web site:

Microsoft Network Monitor 3.4

(http://www.microsoft.com/en-us/download/details.aspx?id=4865)

For more information about the Network Monitor utility, click the following article number to view the article in the Microsoft Knowledge Base:

933741

(http://support.microsoft.com/kb/933741/ )

Information about Network Monitor 3

To use the Network Monitor utility to capture network traffic, follow these steps.

Note The following steps are based on Network Monitor 3.2.1303.0.

1. Set up the Network Monitor utility. To do this, follow these steps:
   1. Start the Network Monitor utility.
   2. Under Select Networks, click to select theLocal Area Connection check box.
   3. Click New Capture.
   4. On the Tools menu, clickOptions.
   5. In the Options dialog box, click theCapture tab.
   6. Under Temporary capture file, change theSize setting from 20 to30.
   7. Click OK.
2. When you are ready to reproduce the problem, ping the client from the server. To do this, type the following command at a command prompt, and then press ENTER:
   ping <ClientName>
3. In the Network Monitor utility, click Start to start the capture.
   
   Note In the HH:MM:SS format, note the time when you start the trace.
4. Start the applications that are typically running on the server.
5. When this problem occurs, note the time in the HH:MM:SS format.
6. Stop the applications.
7. In the Network Monitor utility, click Stop.
8. Save the trace. To do this, follow these steps:
   1. On the File menu, clickSave As.
   2. In the File name box, specify a name in the following format:
      ProblemDescription_ ServerName_ UserName_ HHMM_ MMDDYear
      Note The placeholder ProblemDescription represents a brief description of the problem. The placeholderServerName represents the name of the server that encounters the problem. The placeholderUserName represents the account that you use to log on the server. The placeholderHHMM represents the time when you stop the trace. The placeholder MMDDYear represents the date when you stop the trace.
   3. Click Save.

Important You should also collect the IP address of the server and the IP address of the client.

How to optimize the network load that the clients create

After you analyze both the kind of load on the server and the network trace that you capture, you can optimize the network load that the clients create. To do this, fine-tune the Registry to reduce the network connections to the File Share resources on the server.

After you analyze both the kind of load on the server and the network trace that you capture, you can optimize the network load that the clients create. To do this, fine-tune the Registry on the client computers to reduce the network connections to the File Share resources on the server. The client computers you must do this on include the following:

• Windows XP
• Windows Vista
• Windows 7
• Windows Server 2003 Terminal Server
• Windows Server 2008 Terminal Server

To fine-tune the registry, use one of the following methods.

Important This section, method, or task contains steps that tell you how to modify the registry. However, serious problems might occur if you modify the registry incorrectly. Therefore, make sure that you follow these steps carefully. For added protection, back up the registry before you modify it. Then, you can restore the registry if a problem occurs. For more information about how to back up and restore the registry, click the following article number to view the article in the Microsoft Knowledge Base:

322756

(http://support.microsoft.com/kb/322756/ )

How to back up and restore the registry in Windows

Method 1

1. Start Registry Editor.
   • In Windows Vista, in Windows 7, and in Windows Server 2008, click Start
     Collapse this image Expand this image

     
     , type regedit in the Start Search box, and then press ENTER.
     
     
     Collapse this image Expand this image

     
     If you are prompted for an administrator password or for confirmation, type the password or provide confirmation.
   • In Windows XP and in Windows Server 2003, click Start, clickRun, type regedit in theOpen box, and then click OK.
2. In Registry Editor, locate and then click the following registry subkey: HKEY_LOCAL_MACHINE\Software\Microsoft\Windows\CurrentVersion\policies\explorer
3. Create a NoRemoteRecursiveEvents registry entry, and then set the value to 1 (hexadecimal). To do this, follow these steps:
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type NoRemoteRecursiveEvents, and then press ENTER.
   3. On the Edit menu, clickModify .
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base , type1 in the Value data box, and then clickOK.
4. Create a NoRemoteChangeNotify registry entry, and then set the value to 1 (hexadecimal). To do this, follow these steps:
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type NoRemoteChangeNotify, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type1 in the Value data box, and then clickOK.
5. In HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\MRxSmb\Parameters Create a InfoCacheLevel registry entry, and then set the value to A (hexadecimal). To do this, follow these steps:
   1. On the Edit menu, point toNew, and then click DWORD Value.
   2. Type InfoCacheLevel, and then press ENTER.
   3. On the Edit menu, clickModify.
   4. In the Edit DWORD Value dialog box, clickHexadecimal under Base, type10 in the Value data box, and then clickOK.
6. Exit Registry Editor.

Method 2

1. Start Notepad.
   • In Windows Vista, in Windows 7, and in Windows Server 2008, click Start
     Collapse this image Expand this image

     
     , type notepad.exe in the Start Search box, and then press ENTER.
   • In Windows XP and in Windows Server 2003, click Start, clickRun, type notepad.exe in theOpen box, and then click OK.
2. Copy the following text, and then paste it into Notepad:

   Windows Registry Editor Version 5.00 
   [HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\policies\explorer]
   "NoRemoteRecursiveEvents"=dword:00000001
   "NoRemoteChangeNotify"=dword:00000001
   [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\MRxSmb\Parameters]
   "InfoCacheLevel"=dword:000000010

3. Save the file as .reg file type, and then exit Notepad.
4. In Windows Explorer, double-click the file to import these registry settings