EMC Networkers 'nsrexecd' killed with a segfault

This document (7017650) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 12.x
SUSE Linux Enterprise Server 15.x

Situation

When EMC's NetWorker is executed, nsrexecd is killed with a protection fault.

The message printed looks like:

p, li { white-space: pre-wrap; }

2016-04-18T11:57:56.919597+02:00 XXXX kernel: [439981.930169] traps: nsrexecd[30980] general protection ip:7ff391427198 sp:7ff386ffcaa8 error:0 in libpthread-2.19.so[7ff391416000+18000]

Resolution

Open /etc/ld.so.conf with an editor and add the following line at the beginning.
/lib64/noelision

When done, please save the file and run /sbin/ldconfig.
To verify all is correct run /usr/bin/ldd `which nsrexecd`. It should print something like:

XXXX:~ # /usr/bin/ldd `which nsrexecd`

   [...]

   libpthread.so.0 => /lib64/noelision/libpthread.so.0 (0x00007fc0a9afc000)

   [...]

Cause

A change in glibc exposed a bug in the microcode of the CPU or the crashing application.

There are 2 situations where this work-around is required:

1) Intel CPU models with broken TSX microcode

The CPU models affected by this are/were limited. Intel should have resolved this with a microcode update.

2) Software that relies on undefined behavior when a lock is destroyed twice.

ISVs should fix their software to prevent this from occurring.

Disclaimer

This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.