Changelog for sbd-devel-1.5.1+20211116.6bb085f-116.3.x86_64.rpm :

* Thu Nov 25 2021 Johannes Segitz - Added hardening to systemd service(s) (bsc#1181400). Added patch(es):
* harden_sbd.service.patch
* harden_sbd_remote.service.patch
* Tue Nov 16 2021 Yan Gao - Update to version 1.5.1+20211116.6bb085f (v1.5.1):- configure: validate configure options for paths (bsc#1185182)- man: refer to the modern run state directory `/run` if appropriate (bsc#1185182)- configure: add --with-runstatedir option (bsc#1185182)
* Tue Oct 19 2021 Yan Gao - Update to version 1.5.0+20211005.5ed9fd2:- sbd-md: properly destroy io-context- sbd-md: properly free one-time-allocations of sector-buffers- avoid using deprecated valloc & frequent aligned alloc
* Thu Aug 12 2021 Yan Gao - Update to version 1.5.0+20210720.f4ca41f:- sbd-inquisitor: Implement default delay start for diskless sbd (bsc#1189398)- sbd-inquisitor: Sanitize numeric arguments
* Thu Aug 12 2021 Yan Gao - Update to version 1.5.0+20210629.1c72cf2:- sbd-inquisitor: tolerate and strip any leading spaces of command line option values (bsc#1187547)- sbd-inquisitor: tell the actual watchdog device specified with `-w` (bsc#1187547)
* Thu Aug 12 2021 Yan Gao - Update to version 1.5.0+20210614.d7f447d (v1.5.0):- spec, configure.ac: default to resource-syncing with pacemaker
* Mon May 03 2021 Peter Varkoly - Deprecated path \"/var/run/\" used in systemd-services (bsc#1185182)
* Tue Mar 09 2021 Yan Gao - Update to version 1.4.2+20210305.926b554:- sbd-inquisitor: take the defaults for the options set in sysconfig with empty strings (bsc#1183259)
* Tue Mar 09 2021 Yan Gao - Update to version 1.4.2+20210305.57b84b5:- sbd-inquisitor: prevent segfault if no command is supplied (bsc#1183237)
* Thu Mar 04 2021 Yan Gao - Update to version 1.4.2+20210304.488a5b9:- sbd-inquisitor,sbd-md: make watchdog warning messages more understandable (bsc#1182648)- sbd-inquisitor: calculate the default timeout for watchdog warning based on the watchdog timeout consistently (bsc#1182648)- sbd-inquisitor: ensure the timeout for watchdog warning specified with `-5` option is respected (bsc#1182648)- sbd-common: ensure the default timeout for watchdog warning is about 3/5 of the default watchdog timeout (bsc#1182648)
* Fri Jan 29 2021 Yan Gao - sbd-inquisitor: downgrade the warning about SBD_SYNC_RESOURCE_STARTUP to notice (bsc#1180966)
* bsc#1180966-0001-Log-sbd-inquisitor-downgrade-the-warning-about-SBD_S.patch
* Fri Jan 29 2021 Yan Gao - Update to version 1.4.2+20210129.5e2100f:- Doc: adapt description of startup/shutdown sync with pacemaker
* Thu Jan 21 2021 Yan Gao - Update to version 1.4.2+20201214.01c18c7:- sbd-inquisitor: check SBD_SYNC_RESOURCE_STARTUP only in watch mode (bsc#1180966)
* Fri Dec 04 2020 Yan Gao - Update to version 1.4.2+20201202.0446439 (v1.4.2):- ship sbd.pc with basic sbd build information for downstream packages to use
* Tue Nov 10 2020 Yan Gao - Update to version 1.4.1+20201105.507bd5f:- sbd: inform the user to restart the sbd service (bsc#1179655)
* Fri Oct 30 2020 Franck Bui - Update the uses of the systemd rpm macros
* use \'%service_del_postun_without_restart\' instead of \'%service_del_postun -n\'
* drop use of \'%service_del_preun -n\' as \'-n\' is unsafe and is deprecated This part still needs to be reworked as leaving services running why their package has been removed is unsafe.
* Thu Aug 27 2020 Yan Gao - Update to version 1.4.1+20200819.4a02ef2:- sbd-pacemaker: stay with basic string handling- build: use configure for watchdog-default-timeout & others
* Thu Aug 27 2020 Yan Gao - Update to version 1.4.1+20200807.7c21899:- sbd-inquisitor: refuse to start if any of the configured device names is invalid (bsc#1174915)
* Thu Aug 27 2020 Yan Gao - Update to version 1.4.1+20200727.1117c6b:- make syncing of pacemaker resource startup configurable- sbd-pacemaker: sync with pacemakerd for robustness
* Thu Aug 27 2020 Yan Gao - Update to version 1.4.1+20200727.971affb:- sbd-cluster: match qdevice-sync_timeout against wd-timeout- Rebase:
* bsc#1140065-Fix-sbd-cluster-exit-if-cmap-is-disconnected.patch
* Thu Aug 27 2020 Yan Gao - Update to version 1.4.1+20200624.cee826a:- sbd-pacemaker: handle new no_quorum_demote (rh#1850078)
* Mon Aug 03 2020 Yan Gao - Update to version 1.4.1+20200306.9571749:- scheduling: overhaul the whole thing (bsc#1143064)
* Mon Aug 03 2020 Yan Gao - Update to version 1.4.1+20200130.7e958a9:- Doc: add environment section to man-page
* Mon Jan 13 2020 Yan Gao - Update to version 1.4.1+20200113.4b617a1:- agent: correctly compare string values when calculating timeout (bsc#1148236)
* Mon Jan 13 2020 Yan Gao - Update to version sbd-1.4.1+20191119.25fce8a (v1.4.1):- regressions.sh: relaxed timeouts for tests under load
* Wed Oct 30 2019 Yan Gao - Update to version 1.4.0+20191029.695f9ca:- tests: add regression-tests using preload-library- tests: added preload-library for reboot interception
* Wed Oct 30 2019 Yan Gao - Update to version 1.4.0+20191028.3f01a1d:- spec: add devel package
* Wed Oct 30 2019 Yan Gao - Update to version 1.4.0+20191028.d937f9d:- sbd-inquisitor: use crashdump timeout- Build: switch back to serial test-harness- Doc: mention crashdump message in usage note- defaults: make 15s timeout default for s390 consistently- cmdline: just use SBD_DEVICE if no devs from cmdline
* Mon Oct 28 2019 Yan Gao - Update to version 1.4.0+20190919.2758632:- sbd.8.pod and -h help text add -vvv description- sbd-md: add a warning log if failed to open/read device on startup (bsc#1150429)- agent: log detailed errors for monitor failures (bsc#1148236)- sbd-md: list/dump failures go to stderr (bsc#1148236)- avoid deprecated names for g_main-loop-functions- sbd-pacemaker: check for shutdown attribute on every cib-diff
* Tue Jul 02 2019 Yan Gao - Update to version 1.4.0+20190612.398628b:- sbd-cluster: periodically check corosync-daemon liveness- sbd-pacemaker: assume graceful exit if leftovers are unmanged- sbd-common: query rt-budget > 0 otherwise try moving to root-slice (bsc#1143064)- Rebase:
* bsc#1140065-Fix-sbd-cluster-exit-if-cmap-is-disconnected.patch
* Tue Jul 02 2019 jtomasiakAATTsuse.com- sbd-cluster: Fix 100% CPU usage when CMAP connection is lost (bsc#1140065, SOC-8774)
* bsc#1140065-Fix-sbd-cluster-exit-if-cmap-is-disconnected.patch
* Tue May 14 2019 Yan Gao - Update to version 1.4.0+20190514.e9be8d9:- sbd-inquisitor: Avoid flooding logs with messages that hint the default/configured timeout action (bsc#1134496)
* Tue Apr 30 2019 Yan Gao - Update to version 1.4.0+20190416.5e3283c:- sbd-inquisitor: overhaul device-list-parser- sbd-inquisitor: free timeout action on bail out- sbd-md: prevent unrealistic overflow on sector io calc
* Tue Apr 02 2019 Yan Gao - Update to version 1.4.0+20190326.c38c5e6:- sbd-pacemaker: bail out of status earlier- sbd-pacemaker: make handling of cib-connection loss more robust
* Mon Mar 11 2019 Yan Gao - Update to version 1.4.0+20190311.0159a3c:- sbd-cluster: finalize cmap connection if disconnected from cluster (bsc#1128059)
* Tue Feb 12 2019 ygaoAATTsuse.com- Update to version 1.4.0+20190201.f949aa8:- fail earlier on invalid servants
* Thu Jan 24 2019 ygaoAATTsuse.com- Update to version 1.4.0+20190123.1829c40:- sbd.sysconfig: watchdog timeout set in the on-disk metadata takes precedence- sbd.8.pod: use the generic term \"cluster services\" instead of the specific \"openais\" (bsc#1112918)
* Thu Jan 17 2019 ygaoAATTsuse.com- Update to version 1.4.0+20190114.7f33d1a (v1.4.0)
* Wed Jan 09 2019 ygaoAATTsuse.com- Update to version 1.3.1+20181218.dbae8ab:- make timeout-action executed by sbd configurable- use pacemaker\'s new pe api with constructors/destructors
* Thu Oct 04 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180920.bf16663:- sbd-common: avoid statting potential links- sbd-inquisitor: SBD_DELAY_START can be configured with a delay value (bsc#1107321)- sbd-common: don\'t follow symlinks outside /dev for watchdog
* Fri Jul 27 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180703.f95b98e (bsc#1102930)
* Tue May 08 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180507.e102d9e:- sbd-md: return error if faied to list any devices (bsc#1086650)- sbd-cluster: let scan do the job of proc-parsing- sbd-cluster: search for pacemaker-remoted
* Wed May 02 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180502.970d913:- Log: upgrade important messages and downgrade unimportant ones (bsc#1091839)- Log: change sbd\'s default logging level to LOG_NOTICE (bsc#1091839)
* Sat Mar 17 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180316.8873890:- Fix: add Documentation value to systemd services
* Tue Mar 13 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180313.fe5fe1d:- Build: sbd-pacemaker: Query CIB directly with the API instead of get_cib_copy() (bsc#1084689)
* Mon Feb 05 2018 jengelhAATTinai.de- Combine %service_
* calls to reduce amount of generated code.- Use %make_install, and make sure directories are owned.
* Mon Feb 05 2018 ygaoAATTsuse.com- Update to version 1.3.1+20180202.cc3e4cb:- Doc: sbd.8.pod: add sections for query-watchdog & test-watchdog- sbd-md: dump_headers returns 0 even open_device failed (bsc#1079316)
* Sat Dec 23 2017 ygaoAATTsuse.com- Update to version 1.3.1+20171220.1e93740:- Fix: systemd: make pacemaker & dlm wait for sbd-start to complete (bsc#1108393)- Doc: sbd.sysconfig: mention timeout caveat with SBD_DELAY_START (bsc#1074038)
* Fri Dec 01 2017 ygaoAATTsuse.com- Update to version 1.3.1+20171130.ecbdf9c:- Build: cluster-servant: Compile with pacemaker-2.0
* Wed Nov 29 2017 ygaoAATTsuse.com- Fix: build error with glibc 2.25- Fix: gcc format string error- Upstream version cs: a7a7d792b9922f19bbace276c8a154e157a4f99c
* Thu Nov 23 2017 rbrownAATTsuse.com- Replace references to /var/adm/fillup-templates with new %_fillupdir macro (boo#1069468)
* Fri Nov 03 2017 ygaoAATTsuse.com- Bump to 1.3.1- Add commands to test/query watchdogs- Upstream version cs: a74b4d25a3eb93fe1abbe6e3ebfd2b16cf48873f (v1.3.1)
* Fri Nov 03 2017 ygaoAATTsuse.com- man: Call \"-P\" option \"Pacemaker integration\" in the description (bsc#1033600)- sbd-inquisitor: Do not use watchdog if the conflicting options \"-W\" and \"-w /dev/null\" are both supplied (bsc#1047376)
* Drop bug-1047376_sbd-not-use-watchdog-conflicting-options.patch which has been merged upstream- sbd-inquisitor: Enable pacemaker integration by default as advertised in sysconfig.sbd (bsc#1046421, bsc#1033600)
* Drop bug-1046421_sbd-enable-pacemaker-integration-by-default.patch which has been merged upstream- sbd-inquisitor: Respect the obsolete option SBD_WATCHDOG for backward compatibility (bsc#1047372)
* Drop bug-1047372_sbd-SBD_WATCHDOG-compatibility.patch which has been merged upstream- Upstream version cs: 681ce1acbda12754eb97fb91f9a7306ff368452f
* Fri Nov 03 2017 ygaoAATTsuse.com- cluster-servant: check for corosync 2Node mode- disk-servant: signal reset request via exit-code- improve reboot after sysrq_trigger- Upstream version cs: 8a0800cc906d7bb8f8ac0d5664bdf691b2ca7690
* Fri Jul 21 2017 ygaoAATTsuse.com- Fix node name parameter in manpage (bsc#963674)- Don\'t reboot if off is requested- sbd-inquisitor: cleanup_servant_by_pid() should be static
* Drop sbd-cleanup_servant_by_pid-static.patch which has been merged upstream- sbd.sh: Use a more obvious variable on parsing devices (bsc#1033934)- sbd-inquisitor: Do not create duplicate servants (bsc#1033934, bsc#1065748)- sbd-inquisitor: Correctly look up servant by device name (bsc#1033934, bsc#1065748)- Upstream version cs: 11262fbc80364f2993a9876a2be82f4170c41e4e
* Tue Jul 04 2017 ygaoAATTsuse.com- sbd-inquisitor: Do not use watchdog if the conflicting options \"-W\" and \"-w /dev/null\" are both supplied (bsc#1047376)
* bug-1047376_sbd-not-use-watchdog-conflicting-options.patch- sbd-inquisitor: Enable pacemaker integration by default as advertised in sysconfig.sbd (bsc#1046421, bsc#1033600)
* bug-1046421_sbd-enable-pacemaker-integration-by-default.patch- sbd-inquisitor: Respect the obsolete option SBD_WATCHDOG for backward compatibility (bsc#1047372)
* bug-1047372_sbd-SBD_WATCHDOG-compatibility.patch
* Fri Jun 02 2017 ygaoAATTsuse.com- sbd-inquisitor: cleanup_servant_by_pid() should be static
* sbd-cleanup_servant_by_pid-static.patch- sbd.sh: Correctly handle SBD_DELAY_START for multiple SBD devices (bsc#1033934)- sbd.sh: consistent use of SBD_BIN (bsc#1033934)- Upstream version cs: a0bc43b17fcb997fe4c09731db2126ab350cfb50
* Wed May 31 2017 ygaoAATTsuse.com- Update to v1.3.0- spec: Do not automatically try to restart sbd.service on update (bsc#1033934)- Upstream version cs: 4968e9c8602fbb990bed63cc96ca18f62e2181db (v1.3.0)
* Mon Mar 06 2017 ygaoAATTsuse.com- Clarify documentation around watchdog option (-W)- Fix node-name handling in structs - bail out if too long- man: Update man page to use pcmk_delay_max over start-delay (bsc#993032)
* Drop bug-993032_sbd-man-pcmk_delay_max.patch which has been merged upstream- use LDADD instead of LDFLAGS to fix build failure with ld --as-needed
* Drop obsolete add-explicit-libs.patch- Set SBD_STARTMODE to \"always\" by default- Upstream version cs: 4ee36fa33b57a1ad95678363911cfb02a92e1edb
* Fri Feb 24 2017 ygaoAATTsuse.com- Correctly find the pacemaker remote process- Pacemaker remote must be allowed to start in parallel to sbd- Only log \'healthy\' when a servant was previously outdated- Ensure the correct callbacks are always configured- Streamline the initial cluster connection- Non-disk servants start off as outdated until we hear from them- Correctly check if the new servant is a disk- Ensure crm_system_name is set to avoid SEGFAULT when connecting to the cluster- Repair diskless startup (fate#321017)- Enable the cluster check module- systemd: Provide sbd_remote.service file for pacemaker remote nodes (bsc#962287)
* Drop sbd_remote.service which has been merged upstream- Simplified check for disk vs. non-disk servants- Implement a new mode that gives priority to the cluster servants- Support pacemaker remote as a cluster connect type that we can monitor- Attempt to reconnect to the cluster after a failure- Only continue tickling the watchdog without quorum if we have non-disk elements- Notify the parent explicitly when the cluster connection goes away- Track the number of servants and disks differently- Cluster servant focuses exclusively on having a live cluster connection- Automatically scale timeout_watchdog_warn with timeout_watchdog- Differentiate between when a node is offline and unknown- Gracefully handle an empty CIB- Repair startup when no disks are present (fate#321017)- Allow the watchdog timeout to be specified via the environment- Self terminate when quorum is lost and the admin configured no-quorum-policy=suicide (bsc#950415)
* Drop obsolete bug-950415_sbd-pacemaker-segfault.patch- Prevent watchdogdev variable from being erased on fork()- Fix the parsing of devices from the environment- Let systemd own where the pidfile lives- Use the pacemaker corefile directory- Repair the SBD_STARTMODE handling and allow startup without a disk partition (fate#321017)- Batch up the processing of cib updates but cap the maximum delay- Simplify the systemd unit file by reading options from the environment- Drop dependancy on clplumbing
* Drop obsolete sbd-configure-libcoroipcc.patch- Pacemaker: Fix compilation when building against corosync 2.x
* Drop obsolete sbd-pacemaker.patch- Upstream version cs: c511b0692784a7085df4b1ae35748fb318fa79ee
* Wed Aug 31 2016 ygaoAATTsuse.com- man: Update man page to use pcmk_delay_max over start-delay (bsc#993032)
* bug-993032_sbd-man-pcmk_delay_max.patch
* Wed Jul 20 2016 ygaoAATTsuse.com- spec: Update the Url of the project
* Thu Jun 30 2016 tchvatalAATTsuse.com- Cleanup a bit with spec-cleaner- Use pgkconfig dependencies (and only those that are really needed)- Remove systemd conditions as the only distro versions we build against have systemd- Disable the werror cflags addition, we always build without it distro wide
* Tue May 17 2016 kgronlundAATTsuse.com- Update to version 1.2.1.git.1463482437.d6bd23a:
* Add support for diskless sbd mode- Remove outdated patches:
* Remove sbd-pacemaker.patch
* Remove sbd-configure-libcoroipcc.patch
* Remove bug-950415_sbd-pacemaker-segfault.patch
* Tue Jan 19 2016 ygaoAATTsuse.com- systemd: Provide sbd_remote.service file for pacemaker remote nodes (bsc#962287)
* Thu Oct 15 2015 ygaoAATTsuse.com- pacemaker: Prevent potential segfault caused by use-of-NULL on checking node state (bsc#950415)
* bug-950415_sbd-pacemaker-segfault.patch
* Wed Jun 24 2015 lmbAATTsuse.com- Trigger rebuild of sbd against more recent pacemaker packages (no code change) (bnc#934919)
* Fri Sep 05 2014 lmbAATTsuse.com- Update to version 1.2.1.git.1409904429.39dee2a: + Medium: adjust timeout for s390(x) to be compatible with vmwatchdog (bnc#895103)
* Fri Aug 15 2014 lmbAATTsuse.com- Update to version 1.2.1.git.1408100832.f0a4a63: + systemd: Start sbd before corosync instead of pacemaker (bnc#892078) + systemd: allow delaying start of post-sbd units (bnc#888935)- Update _service file.
* Thu Jun 12 2014 lmbAATTsuse.com- Add _service to pull from latest git.
* Thu Jun 05 2014 lmbAATTsuse.com- pacemaker: improve XML performance and avoid crash with pacemaker >= 1.1.12 (bnc#881231)- man: correct syntax in example for the CIB configuration- cs: 6f717f5
* Fri May 23 2014 jsegitzAATTnovell.com- added necessary macros for systemd files
* Thu Mar 27 2014 lmbAATTsuse.com- sbd: Allow the watchdog to be explicitly disabled (bnc#865365)- rpm: use the short git hash in the changelog from now on.- cs: b96ac28
* Thu Mar 27 2014 lmbAATTsuse.com- systemd: ensure that sbd is activated after systemd module loading and the iSCSI initator (bnc#869612)- Start-up: make sbd wait a configurable amount of time for devices to appear on boot (bnc#869612)- Downgrade the logging during the initial wait to a debug log to avoid log flooding.- cs: 512b71dc209744e505bf62d7ccddb70854fe7407
* Thu Oct 10 2013 lmbAATTsuse.com- systemd support- /etc/sysconfig/sbd template added- sbd: Add -p option to specify a pidfile- sbd-pacemaker: detect whether to run the AIS plugin quorum check at runtime- cs: d7df351e09f903be79495dbe312cf670f1747ef2
* Tue Oct 01 2013 lmbAATTsuse.com- License update in source files and spec file- agent: detect if the stonith-timeout is too low (bnc#841010)- agent: auto-correct pacemaker configuration if a too short timeout is detected. This can be disabled using the timeout_bypass option. (bnc#841010)- man: explain why stonith-timeout needs to be larger than msgwait- Drop fix-check-lib-lib-order.patch: merged upstream- cs: 3757c4f6bc80146077cb2e019537c44d288f414d
* Fri Sep 13 2013 lmbAATTsuse.com- Add copyright comment to spec file.
* Tue Jul 30 2013 ygaoAATTsuse.com- Added sbd-pacemaker.patch: Remove use of ais_fd_sync which only exists in corosync-1.x-based pacemaker plugin- Added sbd-configure-libcoroipcc.patch: No longer check libcoroipcc when configuring- Added add-explicit-libs.patch: add explicit libraries to build process.
* Thu Jul 04 2013 lmbAATTsuse.com- sbd-pacemaker: Improve performance- sbd-pacemaker: handle transient failures of the CIB update processing- sbd-pacemaker: Log two more cases where pacemaker state could become unhealthy- Bump version number to 1.2.0- cs: ffca2d93872b81f8b10ef1164c024005576afd4b
* Fri Jun 28 2013 lmbAATTsuse.com- sbd-pacemaker: update to support build with pacemaker-1.1.10
* Wed May 22 2013 tserongAATTsuse.com- Fix library link order, add required libraries explicitly + fix-check-lib-lib-order.patch + add-explicit-libs.patch
* Mon Mar 25 2013 lmbAATTsuse.com- Introduce -S option to prevent automatic start if the node was previously fenced (bnc#812195)- Reduce number of scenarios where usage help is printed (bnc#812191)- cs: 1837fd8cc64a
* Wed Feb 27 2013 lmbAATTsuse.com- Introduce minor header version and add a UUID (bnc#804991)- Add debug mode level 3 to allow logs to flush, but do actually reboot.- cs: 9641481ebc9b
* Thu Nov 22 2012 lmbAATTnovell.com- Failure to set watchdog interval is now fatal (bnc#790894)- cs: d262617db1eb
* Thu Nov 15 2012 lmbAATTnovell.com- Upstream merge of compatibility changes for pacemaker 1.1.8- cs: 52ddf89aa96b
* Mon Oct 22 2012 ygaoAATTsuse.com- Changes for being compatible with pacemaker >= 1.1.8
* Fri Sep 28 2012 lmbAATTnovell.com- Increase sbd\'s scheduling priority to the maximum (bnc#779259)- Print proper defaults for -t/-F in sbd help (manpage was already correct)- cs: 6689be9d3aec
* Mon Jun 18 2012 lmbAATTnovell.com- Remove explicit conflict with cluster-glue.
* Mon Jun 18 2012 lmbAATTnovell.com- Include a man page for the sbd command.- Conflict with older versions of ClusterTools2 due to a file conflict.- cs: 3a371cd0f200
* Tue Jun 12 2012 lmbAATTnovell.com- cs: 68e33d323c6e- Include more detail on the result of failed async IO in log- Fix access to unallocated memory in servant
* Wed May 30 2012 lmbAATTnovell.com- cs: 7d7e0803def7- Log once if pacemaker state prevented self-fence due to lost device majority- Build using autoconf/autotool
* Tue May 29 2012 lmbAATTnovell.com- cs: d7c676111472- Make write IO asynchronous as well.- Ensure that each mbox will only be zeroed once.- Initialize sysrq and enable reboot/poweroff flags.- Unify async IO read/write paths.
* Fri May 25 2012 lmbAATTnovell.com- Make the code compile on SP1 too, though AIS integration is not available there.
* Fri May 25 2012 lmbAATTnovell.com- cs: dcf1f404b6b7- Refresh pacemaker state once per second instead of only on CIB update- Pacemaker state will only be accepted if it is fresh enough- Refetch the full CIB periodically to notice stalls on notifications- Request quorum state also from corosync/AIS, not just the CIB
* Thu May 24 2012 lmbAATTnovell.com- cs: 1cd63d9a962d- Servants should only be considered live if reporting within timeout_io, not timeout_watchdog- Successful delivery should not be logged at ERROR severity
* Thu May 24 2012 lmbAATTnovell.com- cs: 01b157f5cfee- Ensure that sbd latency does not turn negative and thus would cause an unnecessary self-fence.- Drop conflict with older version of cluster-glue; file level conflict is sufficient.
* Tue May 22 2012 lmbAATTnovell.com- cs: 7379080b64ea- Take pacemaker quorum and node state into account for fencing decision (FATE#310625, bnc#767293)- Adjust sbd servant restart intervals- Allow setting a different watchdog timeout before crashdump (bnc#762300)- Handle loss of connection to the CIB process properly- Split sbd off cluster-glue- Limit number of no restart messages printed (bnc#753559)- Actually don\'t restart servants if restart limit reached- Add debug mode via -Z commandline switch (bnc#753559)- Fix allocation in the face of failures (bnc#753559)- handle stuck IO better by explicitly informing our parent 738295)- Make servant restart logic more robust and verbose (bnc#738295)- Make sure the restart limit set to zero works (bnc#738295)- Use async io for reads (bnc#738295)- Skip missing/broken devices during list/dump (bnc#761797)- Add strategic log messages for debuggability