000023685 - aceserver_fe process crashes frequently on ACE/Server 5.2.1

Document created by RSA Customer Support Employee on Jun 14, 2016Last modified by RSA Customer Support Employee on Apr 21, 2017
Version 2Show Document
  • View in full screen mode

Article Content

Article Number000023685
Applies ToACE/Server 5.2.1
Solaris 2.8
_aceserver_fe
_aceserver_be
Issueaceserver_fe process crashes frequently on ACE/Server 5.2.1

The front end  of authentication, _aceserver_fe process crashes frequently either on primary or replica server

pstack output:
ff133c7c strncpy (f9d01858, 0, fb, 0, 0, f9d01858) + 5b4
000d0018 errSDGetHostByAddr (768f88, f9d01858, f9d01828, 100, 30, ff29ef68) + f4
000cfe4c errSDGetFQNameandIPAddr (770a28, f9d01858, f9d01828, 100, 30, 0) + 130


mdb analysis of core file:

mdb _aceserver_fe <filename>.core
Loading modules: [ ]
> $C
f9d01580 libc.so.1`strncpy+0x5b4(f9d01858, 0, fb, 0, 0, f9d01858)
f9d015c0 errSDGetHostByAddr+0xf4(768f88, f9d01858, f9d01828, 100, 30, ff29ef68)
f9d01720 errSDGetFQNameandIPAddr+0x130(770a28, f9d01858, f9d01828, 100, 30, 0)

Cause

The host routine errSDGetFQNameandIPAddr() was being called from the lock manager which is multi-threaded. The errSDGetFQNameandIPAddr() routine is not thread safe which caused the Ace Server front end to core.

Resolution

This has been identified as BugID56057 and a fix is available to resolve  this issue. Please contact RSA Security Technical support to obtain this fix. The lock manager was changed to call the errSDGetFQNameandIPAddr() as a non-thread safe routine. The fix contains _aceserver_fe and _aceserver_be files. These two files should be replaced on Primary server and all replica servers in ace/prog directory.

1. Stop the RSA ace services ( aceserver stop, sdconnect shutdown).

2. Replace the binaries from the hot fix tar file in ace/prog directory.

3. Start RSA ACE services ( sdconnect start, aceserver start)

Legacy Article IDa35034

Attachments

    Outcomes