VCS engine (had) restarts due to segmentation fault (SIGSEGV), also consumes lot of memory
Article ID: 100025962
Updated On:
Description
Error Message
HAD core dumped as it was unable to allocate memory as shown in /var/adm/messages:
Jan 22 13:59:45 node01 Had[17579]: [ID 702911 daemon.alert] malloc() FAILED at file VCSStrLib.C, line 106, memory could not be allocated. VCS engine dumping core deliberately!
Jan 22 13:59:45 node01 Had[17579]: [ID 702911 daemon.alert] VCS WARNING V-16-1-53032 HAD Signal SIGSEGV received
Threadlist of had core shows the following stack:
17579: /opt/VRTSvcs/bin/had
----------------- lwp# 1 / thread# 1 --------------------
ff14ccb4 waitid (0, 5118, ffbee178, 3)
[...]
--- called from signal handler with signal 11 (SIGSEGV) ---
0028fd98 __1cJvcsmalloc6FIpcI_pv_ (4000, 38cab0, 6a, 0, 0, 0) + 118
0028fa24 __1cLvcssnprintf6FpcIpkcE_I_ (fe8f8b28, 4000, 354e64, 0, 80808080, 1010101) + 5c
[...]
0008a654 main (1, ffbffa34, ffbffa3c, 3d0c00, ff1e0200, 0) + 4c
0006e5c8 _start (0, 0, 0, 0, 0, 0) + 108
Cause
VCS engine core dumped due to shortage of memory. The pmap on the core file shows that the heap size is approximately 4G.
# pmap core
core 'core' of 17579: /opt/VRTSvcs/bin/had
pmap: warning: librtld_db failed to initialize; shared library information will not be available
00010000 3712K r-x-- /opt/VRTSvcs/bin/had
003BE000 344K rwx--
00414000 4166568K rwx--* [ heap ] <== NOTE
..
..
FF3F4000 8K rwx--*
FF3F6000 8K rwx--*
FFBD8000 160K rw---* [ stack ]
total 4177904K
Number of memory leak issues mentioned in the supplemental materials section has been fixed in VCS 5.0 MP3 RP5.
Resolution
Install VCS 5.0 MP3 RP5 to address the memory leak issues.
Applies To
Solaris 10
VCS 5.0 MP3 RP4
Issue/Introduction
Additional Information
