Error Code details
V-16-1-50053
Severity: Critical 
Component: Cluster Server 
Message:
Insufficient memory to read GAB message. Exiting.
Description:

The message is displayed after Global Atomic Broadcast (GAB) detects that the Veritas High Availability Daemon (HAD) exited due to insufficient memory on the server. For example, on the Windows platform, this situation occurs when the server fails to allocate memory from the system’s nonpaged pool because the pool is empty.

 

Veritas solutions
Solution 1 Vote: [Useful] [Not useful]
Last Modified: 2013-02-06 08:09:19 PDT
Platform: Windows Server 2003 (x86-64), Windows Server 2003 (IA-64), Windows Server 2003 (x86-32), Windows Server 2008 (x86-64), Windows Server 2008 (IA-64), Windows Server 2008 (x86-32), Windows Server 2008 R2, Windows Server 2012, Windows Server 2012 R2, Windows Server 2016
Release: Generic
Content:

To determine the cause of the problem and fix it, try any of the following:

 

  • Examine the size of the registry. In some cases, very large registry sizes cause the system to halt. To reduce the size of the registry, remove unwanted data.

 

  • Examine kernel memory usage with tools such as Process Explorer (available at http://live.sysinternals.com). Reduce kernel memory usage by reducing the overall server load, removing unwanted processes, and so on. If you can’t reduce server load, consider adding more memory to the server.

 

  • Examine the system event logs further for information. The issue may be outside of Veritas Cluster Server (VCS). To open Windows Event Viewer, on a command prompt, enter:

start eventvwr

 

  • Identify and analyze any kernel modules that may leak memory, as the memory consumption continues to increase beyond reasonable amounts.  After you identify a module, you can contact the vendor or provider for an update.

 

  • Run Windows Update to get any updates available from Microsoft.

 

  • Configure performance monitor counters on the CPU, memory, and disk.

 

  • If you find the crash dump of a system process or an HAD process, contact Symantec Technical Support.
Solution 2 Vote: [Useful] [Not useful]
Last Modified: 2013-02-06 08:02:12 PDT
Platform: Generic
Release: Generic
Content:

For Unix platforms, to determine the cause of the problem and fix it, try any of the following:

 

  • Identify and analyze any kernel modules that may leak memory. You may need to contact the operating system vendor or check the Internet for any known issues with the kernel.

 

  • Check all the recent driver or patch updates for any known issues regarding memory problems. You may need to check the Internet for related update information.

 

  • Identify any other processes and applications running when the issue happened. If possible, abort the processes or the applications.

 

  • Examine the UNIX syslogs in /var/log/. The issue may be outside of VCS.

 

  • Configure Performance Monitor counters on the CPU, memory, and logical and physical disk.

 

  • If you find the HAD process creates a crash dump under $VCS_DIAG/had.timestamp, contact Symantec Technical Support. The default path is /var/VRTSvcs/diag/had.timestamp.

 

User solutions
No user solutions yet.

 

Read and accept Terms of Service