Update OpenSAF_AMF_PR for the 4.3 release
Update OpenSAF_AMF_PR for the 4.3 release
amf:Updated OpenSAF_AMF_PR.odt for OpenSAF 4.5 [#819]
AMF: Add support to change saAmfSIRank at runtime [#1162]
amf: update for nodegroup lock-instantiate and unlock-instantiate [#1327]
amf: update for nodegroup lock-instantiate and unlock-instantiate [#1327]
clm: Updated CLM Programmers Reference with the new cluster reboot support [#2053]
util: Fenced should only write a log record when two acitve controllers is seen
commit 81b9b1791eb18fda01768d1a5b48763bcb0fd990 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Mon Aug 26 08:18:37 2019 +0200
util: Fenced should only write a log record when two acitve controllers is seen [#3073]
util: Fenced should only write a log record when two acitve controllers is seen
util: Fenced should only write a log record when two acitve controllers is seen [#3073]
util: Fenced should only write a log record when two acitve controllers is seen
util: Write a log record when two acitve controllers is seen
util: Use a fence daemon as an alternative to payload reboot fencing
commit 1d100eca97198639ec04a190c321c464c044b896 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Mon Jun 24 08:40:00 2019 +0200
utils: Use a fence daemon as an alternative to payload reboot fencing [#3048]
util: Use a fence daemon as an alternative to payload reboot fencing
utils: Use a fence daemon as an alternative to payload reboot fencing [#3048]
util: Use a fence daemon as an alternative to payload reboot fencing
mds: use new TIPC getsockopt to log receive buffer utilization
commit 135b0b8862da9a036553c5db02062edb278089aa Author: Hans Nordeback hans.nordeback@ericsson.com Date: Mon Jun 3 13:32:39 2019 +0200
mds: use new TIPC getsockopt to log receive queue utilization [#3038]
mds: use new TIPC getsockopt to log receive buffer utilization
mds: use new TIPC getsockopt to log receive queue utilization [#3038]
mds: use new TIPC getsockopt to log receive queue utilization [#3038]
mds: use new TIPC getsockopt to log receive buffer utilization
mds: Improve tipc receive logic
base: Improve backtrace print in daemon.c
commit 830969232b6db23fa81c0c66f8c3350b5b13e3e8 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Thu Jun 14 11:59:21 2018 +0200 base: Improve backtrace print in daemon.c [#2853]
base: Improve backtrace print in daemon.c [#2853]
mds: return success at failure in case of using broadcast send type
commit 9b8880dc2d05ff561bb2e2be02213a1c04c21417 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Tue May 29 13:14:47 2018 +0200 mds: return success at failure in case of using broadcast send type [#2866] commit 1c4c6ad57fdf44b228ec83f777ae957e878f1d61 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Tue May 29 13:14:47 2018 +0200 mds: return success at failure in case of using broadcast send type [#2866]
mds: return success at failure in case of using broadcast send type [#2866]
mds: return success at failure in case of using broadcast send type [#2866]
mds: return success at failure in case of using broadcast send type [#2866]
mds: return success at failure in case of using broadcast send type
mds: return success at failure in case of using broadcast send type
It seems the daemon_exit() function needs to be modified a bit as you say, just calling _Exit and the cleanup/unlink will not be done.
use daemon_exit() if daemonize() is used, (as is the case in immd, immnd)
base: Destructor of TraceLog causes coredump V2 [#2860]
osaf: Faulty constructs in opensaf detected when compiling with gcc 8.1
mds: Improve tipc receive logic
mds: Improve tipc receive logic
base: Improve backtrace print in daemon.c [#2853]
base: Improve backtrace print in daemon.c
Add support for google gmock framework
commit aff54ff091727f27830443332b830890668749cf Author: Hans Nordeback hans.nordeback@ericsson.com Date: Wed Apr 11 11:56:26 2018 +0200 build: Add support for google gmock framework V2 [#2823]
build: Add support for google gmock framework V2 [#2823]
base: Check return code from unlink in nid_create_ipc
commit b83be452a25a37c7f5b568b436d1af544afb7350 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Mon Apr 9 13:42:28 2018 +0200 base: Check return code from unlink in nid_create_ipc [#2829]
base: Check return code from unlink in nid_create_ipc [#2829]
base: Check return code from unlink in nid_create_ipc [#2829]
base: Check return code from unlink in nid_create_ipc
build: Add support for google gmock framework V2 [#2823]
build: Add support for google gmock framework [#2823]
Add support for google gmock framework
Add support for google gmock framework
Add support for google gmock
base: Improve logging at daemon exit
commit dc467e7e143d113bc11445c909bd8520aed6dfd7 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Tue Mar 20 15:30:15 2018 +0100 base: Improve logging at daemon exit [#2704]
base: Improve logging at daemon exit [#2704]
tools: Fix trace2dot, it stopped working after ticket #2165 [#2668]
tools: trace2dot stopped working after ticket #2165
commit 3139dc2f5f727aa420e9d1cc94450f345b448088 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Fri Mar 16 14:30:09 2018 +0100 tools: Fix trace2dot, it stopped working after ticket #2165 [#2668]
mds: mds test is not thread safe
tools: Fix trace2dot, it stopped working after ticket #2165 [#2668]
tools: trace2dot stopped working after ticket #2165
tools: trace2dot stopped working after ticket #2165
I have not had any time to work on this, so good if you can look at it. Attached a small patch to start with, if you use valgrind helgrind you should see a decrease in remarks. It seems that there are a lot of global data shared between the threads that are not thread safe. There are also memory leaks, check with valgrind memcheck.
and, please feel free to work on these tickets
isn't it better to merge this ticket with #2746? I have not checked this ticket that closely but I assumed it was related to thread safety? Have you run helgrind, (and memcheck)?
we have to run both valgrind memcheck and helgrind to improve the mds test
ticket #2746 is about making mds test thread safe. Running helgrind shows a large number of thread safe issues and after fixing these issues you mean there will be more thread safe issues handled by this ticket?
see ticket https://sourceforge.net/p/opensaf/tickets/2746/ , seems #2798 is a duplicate of this ticket, I have one small part finished but not the complete solution. What do you say, I can assign #2746 to you and send you a patch of what has been done and you to finish the work? /Regards Hans
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf
commit d3145b9c15c64897f46f3232f0c04dbe68a6beaf Author: Hans Nordeback hans.nordeback@ericsson.com Date: Wed Jan 17 12:26:59 2018 +0100 fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf [#2761] commit e1e0d2c0dc45a5ca7789f19d58dde0a41ed19354 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Wed Jan 17 12:26:59 2018 +0100 fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf [#2761]
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf [#2761]
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf [#2761]
commit d082d2fb2b26437bbe6860e8efff8479748378c2 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Tue Jan 16 12:54:55 2018 +0100 build: Add cpplint filter to allow unapproved C++11 header [#2760]
build: Add cpplint filter to allow unapproved C++11 header
build: Add cpplint filter to allow unapproved C++11 header [#2760]
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf [#2761]
fmd: Wrong unit for FMS_NODE_ISOLATION_TIMEOUT in fmd.conf
build: Add cpplint filter to allow unapproved C++11 header [#2760]
build: Add cpplint filter to allow unapproved C++11 header
amfnd: Improve log message for amf_saImmOmAccessorGet_o2
amfnd: Improve log message for amf_saImmOmAccessorGet_o2 [#2755]
amfnd: Improve log message for amf_saImmOmAccessorGet_o2
mds: mds test is not thread safe
mds: mds test is not thread safe
the immd error message will be changed to a warning message. Sending a reply to immnd may fail for several reasons, e.g. the node goes down after sending "introduce-me" messages or the topology message arrives after the send. The immnd will re-send "introduce-me" messages until accepted, so this immd message send failure is not considered as a fatal error.
dtm: Update osaflog.cc to combine a selected number of different logs, (mds, syslog, trace) and merge them into one html file.
dtm: Update osaflog.cc to combine a selected number of different logs, (mds, syslog, trace) and merge them into one html file.
dtm: Update osaflog.cc to combine a selected number of different logs, (mds, syslog, trace) and merge them into one html file.
At failover smfd's vdest is in "transit" and during this time all smfnd has the "MDS Await Active Timer" started. This timer may expire before smfd sets a new vdest and if so the NCSMDS_DOWN msg for the smfd vdest is sent to all smfnd's and the cb->smfd_vdest is set to 0. Then the new active smfd sets the vdest and the message NCSMDS_NEW_ACTIVE is sent to all smfnd's. This message is not handled and leads to all smfnd's still has the cb->smfd_vdest set to 0 and are not able to communicate with the...
commit 3c660eef1ad1ca264e1b143151b33378c8f677d2 Author: Hans Nordeback hans.nordeback@ericsson.com Date: Wed Nov 29 14:44:30 2017 +0100 base: Fix opensaf_scale_out script to handle binary arguments [#2703] commit 3d753c5847af88107a9321601263fb03cff6170f Author: Hans Nordeback hans.nordeback@ericsson.com Date: Wed Nov 29 14:44:30 2017 +0100 base: Fix opensaf_scale_out script to handle binary arguments [#2703]