The Linux-HA team proudly announces bug-fix release 2.1.3 of the Linux-HA (aka "Heartbeat", aka "OpenHA") software.
As usual, you can find it here:
http://linux-ha.org/download/index.html#2.1.3 This release has been extensively tested and is considered stable. At
this time, there are no known regressions from previous stable releases,
or the Novell SLES10 SP1 release.
As usual, any new bugs you find should be documented in bugzilla:
http://old.linux-foundation.org/developer_bugzilla/query.cgi?product=Linux-HA Special thanks to Dejan Muhamedagic, Dave Blaschke and Lars Marowsky-Bree and all the other testers who worked so hard to make sure it was well tested.
It includes some new manageability features, and a few new and improved resource agents and STONITH plugins, and nearly 400 bug fixes and lesser enhancements.
The new features include these:
hb_report: heartbeat problem reporting utility
ciblint: tool for finding CIB errors and anomalies
Added snmp support for many R2 features
OpenBSD compatibility
New resource agents:
iscsi, tomcat, SphinxSearchDaemon, ids, Xen (improvements)
New STONITH support:
external/ipmi, ibmrsa-telnet
Enjoy!!
Change Log:
+ hb_report: heartbeat reporting utility
+ ciblint: new tool to lint a CIB
+
snmp subagent: support for crm + RA iscsi: new OCF resource agent (FATE 301250)
+ RA SphinxSearchDaemon: new OCF resource agent
+
RA tomcat: a new OCF agent + RA ids: a new Informix OCF resource agent
+
stonith ipmi: new external stonith plugin + stonith ibmrsa-telnet: new external stonith plugin
+ Xen RA: new features (memory mgmt, DomU monitor hook)
+ Debian: add libxml2-utils as a dependency: we need xmllint for the BSC
+ pgsql RA: postmaster confusion
+ apache RA: remove children in case the top process didn't
+
ldirectord: add forking mode +
ldirectord: Various updates to OCF compatible RA + ccm (LF 1806): restart all instead of exiting (low impact)
+ High: Heartbeat: Include an option to start the CRM in respawn mode
+ dopd: make it work with multiple concurrent outdate requests
+ dopd: don't hardcode path of drbdadm; it has to be in the PATH, though.
+ dopd: fix potential segfault due to leakage of invalid/freed pointer
+ LF bug 1795 - timeout control for crm_mon --one-shot
+ CTS - if we can't get all the nodes running again, call an external script
+ fix a problem building on Red Hat with tog-pegasus installed
+ pgsql RA: handle the missing pg software properly
+ LF bugzilla 1667: ppc64 RPMs contain 32-bit binaries
+ mgmt: Translations for the new parameters of pengine
+ mgmt: Remove the default settings for master_slave from attlist
+ mgmt: Bugfix: Error caused by blank description of parameter
+ mgmt (LF 1724): Prevent saved login information from being corrupted
+ add bzip2-devel as a build dependency for Fedora and RH
+ fixes for fedora: system(3) function return value cannot be ignored
+ Filesystem: Make mount point required for the GUI
+ dopd: don't cl_free variables on local stack
+ build (LF 1803): add -fgnu89-inline to CFLAGS if gcc version is >=4.1.3 and <4.3
+ haresources2cib.py: set default-action-timeout to the default (20s)
+ pgsql RA: move check for root down to enable meta-data for regular users
+ LF 1678: stonith command core dumps with -h option
+ haresources2cib.py: update ra parameters lists
+ Low: lrmd: Make the -r option the default behavior
+ OpenBSD bugs: 1731 1761 1743 1659
+ CTS: changed a test from causing a reboot to not triggering a reboot
+ CTS: increased timeouts and made the stonithd test use the livedangerously option
+ RPM specfile change: made is_fedora macro turn off is_rhel macro
+ Corrected a warning found on FC8 in the cl_reboot code
+ LF 1794: Modified the external/ssh plugin to have a livedangerously flag
+ LF 1794: external/ssh should first check if the host is reachable
+ RA Filesystem: add lustre to the list of fs which don't need fsck
+ ldirectord: SimpleTCP service check documentation
+ ldirectord: SimpleTCP service check
+ LF Bug 1732: Error handling is bad for comm media
+ High: cib: Make sure everyone uses (add|get)_message_xml() for data fields
+ RA SAPDatabase/SAPInstance: report proper exit codes in case required software is not installed
+
RA mysql: don't look for mysql users in the OS database + High: CRM: Compressed XML needs to be retrieved as a binary blob, not a string
+ High: CRM: Future-proof get_message_xml()
+ High: GUI: Fix compilation after changes to update_attr() form the CIB
+ Low: TE: Logging - unconfirmed actions can happen under normal conditions
+ High: heartbeat: Fix getnodes() compilation with gcc 4.3
+ Medium: PE: Create a syntactic shortcut for the common use-case of "{resource} prefers {node} with {score}"
+ High: CIB: Fix the behavior of update_attr() and delete_attr() when the command is ambiguous
+ ipc (LF 1339): enable message compression
+ Changed HIGHLY to STRONGLY in how strongly we recommend STONITH
+ LF bug 1690: there should be a tool to audit (validate) node names in constraints among other things...
+ Low: clplumbing: Fix log message: size_t != int on 64-bit arch's
+ Low: Admin: crm_standby - Don't complain abot missing values, print the default value instead
+ Low: PE: Update the expected graphs to include the batch-limit option
+ Medium: TE: Allow the CRM to limit the number of resource actions the TE can execute in parallel.
+ Low: crmd: Provide a more informative error when DCs detect other DCs during a join
+ Low: CTS: The Process class requries an integer for triggersreboot
+ High: CRM: In ccm_have_quorum(), an OC_EV_MS_PRIMARY_RESTORED event means we
_do_ have quorum
+ Low: crmd: Cleanup - remove all 'return I_NULL' statements from the FSA
+ High: PE: Set next_role recursivly so that group promotion will work
+ High: PE: Introduce a new API call for determining the location of complex resources
+ Medium: PE: Dont make changes to location constraints, when applying to groups, persistent
+ High: crmd: Prevent shutdown hangs caused by pending ops that can't be cancelled - because the no longer exist in the lrm
+ Medium: Do not disconnect blocking write processes, but discard traffic.
+ Novell 293922: Don't use non-blocking writers, but instead don't block
+ Changed the version of RHEL requiring OpenHPI to be 6 or greater...
+ LF Bug 1662: Put in some RHEL changes provided by Keisuke MORI
+ LF bug 1662: PM spec file should be more usable (Fedora changes)
+ Medium: crm: Dont generate core files for non-fatal assert (ie. the CRM_CHECK macro).
+ Low: PE: ptest - Handle malformed inputs more gracefully
+ Low: crmd: Minor logging improvement to update_dc()
+ Low: crm: Handle a few xml-related error conditions without resorting to CRM_ASSERT
+ Low: Admin: Do a full simulation when we have the live CIB - since we'll always have a status section
+ High: PE: Fix a botched commit (cs: 3de5760b06e0) that incorrectly allowed resources to start without quorum or stonith
+ Medium: PE: Disable delete-then-refresh code. The crmd will remove the resource from the CIB itself making the refresh redundant
+ Medium: PE: All other things being equal, prefer to keep non-failed instances alive
+ Medium: attrd: Bug 1776 - Attrd doesn't exit or reconnect when the CIB is respawned
+ Low: PE: Actions for stonith agents should never default to requiring fencing or quorum
+ High: PE: increment_clone() did not overflow correctly 9->10, 99->100, etc
+ Medium: crmd: Don't remap LRM_OP_PENDING when building full lrm updates
+ Medium: TE: Pending operations shouldn't be processed
+ High: PE: Bug 1765 - Prevent master-master colocation constraints from preventing slaves from starting
+ Low: PE: Reduce logging severity regarding creation of notifications
+ Low: PE: Split the contraint unpacking code into its own file
+ hbagent: reset gMembershipTable on dropping global resources
+ hbagent: fix memory leaks on dropping global resources
+ hbclient: fix memory leaks on signon/signoff
+ RA Xinetd (LF 1742): multiple fixes
+ LRM regression test: make common.filter (sed) script work with non-gnu sed (Darwin)
+ LRM regression test: use LSB_RA_DIR instead of /etc/init.d
+ LF 1766: Heartbeat installs man pages in the wrong directories by default...
+ LF bug 1772: apphbd needs to be able to have clients declare themselves critical resulting in a "fast fail" of the system when they fail
+ debugging for LF bug 1393 - includes a bug fix for the postfork() callback in the GSource.c module for the tempproc trigger Gsource
date: Fri Nov 09 06:54:25 2007 -0700
+ Put in simple debugging code for CCM negative uptime message - plus made some related code slighly safer.
+ HBcomm: use proper define (MSG_DONTWAIT instead of MSG_NONBLOCK)
+ Novell 293922: Heartbeat network processes would block on full buffers.
+ LF bug 1393 - Cannot rename /var/lib/heartbeat/hostcache.tmp to /var/lib/heartbeat/hostcache (scope, risk and severity: all minor)
+ Trying a different approach to get debug info for LF bug 1393 (low risk, important for debugging)
+ Added -p pidfile option to cl_respawn
+ LF bug 1706 (finishing up associated issues)
+ mgmtd (LF 1719): fix the pam file for suse
+ raised timeouts in CTS to account for a slow machine I own
+ LF bug 1705 - cl_respawn core dumps when given -h or --help
+ ccm: speed up ccm considerably in case a node is alone in the membership
+ LF bugzilla 1757: apache resource agent grep methodology can't handle newlines
+ Added debugging and very minor code improvements for LF bug 1393
+ mysql RA (LF 1760): defaults for OpenBSD
+ BSC (LRM test): replace cron with heartbeat as an lsb agent
+ CTS: changed the code to wait for nodes to come up before we try and get uuid list
+ changed configure.in to force 64-bit objects when on gcc-based ppc64 platform
+ Removed RPM dependency on sysklogd - because package name can vary
+ Moved python gtk dependencies into GUI package
+ ccm (LF 1546): ensure that the membership instance number never decrements
+ stonithd: check for storage size when copying the host list
+ stonithd (LF 1727): shuffle dropping privileges around again
+
[LDIRECTORD] allow per-virtual checkinterval configuration + stonithd: drop root privileges earlier
+
Debian: rename ha_logd.cf to logd.cf in debian/heartbeat.files + stonithd (LF 1727): fix dropping privileges
+ Low: heartbeat: Use the correct define when choosing to enable valgrind
+ Medium: crmd: Prevent shutdowns initiated immediately after a node is removed from the cluster with hb_delnode from stalling.
+ [DEBIAN] Add dependacy on gawk as it is required by the OCF IPaddr resource
+ [lvs-users] Patch for ldirectord when using mysql service
+ Low: Fix Makefile.am reference (ha_logd.cf -> logd.cf)
+ Low: Rename example ha_logd.cf file to the proper name.
+ Low: cib: Increase the retry interval when connecting to the ccm to 3s
+ Low: cts: Remove the backup CIB too if --clobber-cib is specified
+ Medium: crmd: Bug 1737 - Inconsistent join state detected - Possible fix
+ High: cib: G_main_del_IPC_Channel() doesn't like being called with a NULL pointer and crashes
+ Low: Fix compile of mgmtd (fixes cs 575ceac23368)
+ Medium: mgmt: Use the new bit-field for resource_t objects
+ Low: cib: Set callback_source=NULL after call to G_main_del_IPC_Channel()
+ Low: crm: Reduce severity of digest mismatch message in apply_xml_diff
+ High: crm: Bug 1749 - Make sure the diff-related digests contain the complete CIB, not just the first line
+ Low: PE: Convert all the resource's boolean flags into a single bit-set
+ ccm (LF 1723): fix logging
+ lrmd (LF 1729): revise return codes
+ Low: heartbeat: Some compilers feel that the fcli variable could be used uninitialized - shut them up
+ Low: core: proctrack - Tweak some log messages to be of a similar form to those in heartbeat.c
+ High: PE: Bug 1712 - Ensure manditory ordering constraints can cause complex resources to be shut down
+ Low: cts: Remove patterns that are no longer of interest
+ Medium: Tools: Bug 1738 - the crmd ignores some requests from crm_resource because it exited too quickly
+ Low: PE: Reduce log priority for some development logging
+ Low: cts: Remove erroneous search pattern from component fail test
+ Low: crmd: Use the unaltered rc for timed out operation events from the lrmd
+ Low: Build: Make sure make clean works in the cts directory
+ Medium: cib: Include digests of the cib a diff was made from and verify it when applying
+ High: PE: Prevent use-of-NULL when the admin creates a colocaiton constraint with an empty group
+ LF bug 1712: ha_logd prevented node from shutting down - changed code to use the waitout() primitive
+ Put in a missing include of <memory.h> for the DRBD code
+ LF bug 1700: HA Signon always reports success
+ rpm build change: allow openhpi to be optional, but default it to be required. Don't require it for SUSE < 10.2
+ stonithd (LF 1727): change attach shmem seg to read only
+ Low: cts: Add a missing comma to the patterns list
+ Low: crmd: Use EXECRA_OK instead of hard-coding rc=0
+ Low: crmd: Increase the retry interval when stalling the FSA to 2s (used when connecting to the ccm, lrmd, cib)
+ Low: crmd: Remove code made redundant by the use of ProcTrack
+ Low: cts: Wait for the cluster to see the node come back before checking for S_IDLE after shooting the ccm.
+ Low: cts: Add an extra ignore pattern for stonithd
+ Low: cts: NearQuorumPoint - only check for Pat:DC_IDLE if there will be nodes up at the end
+ Low: PE: Produce a config error when a clone contains more than one resource/group to clone
+ Low: RA: PureFTPd - Support debian's pure-ftpd-wrapper script
+ Medium: Tools: ocf-tester - Ensure OCF_ROOT is available to the RA subshell
+ Low: contrib: dopd - Remove a pointless and annoying dependancy on the crm
+ Low: TE: Code cleanup
+ stonithd: improve logging
+ LF bug 1502: udpport statement in ha.cf is ignored by heartbeat
+ LF bug 1589: ia64 heartbeat unaligned access messages (SGI965396) - copy fields in structures
+ LF bug 1679: hardcoded heartbeat userid and group in init script
+ LF bug 1681: BSC fails to identify active interface on OpenBSD
+ LF bug 1702: during emergency restarts, heartbeat doesn't close watchdog correctly
+ Updated a specfile dependency to require the minimum version of openhpi
+ LF bug 1734 - removed duplicate return statement as per bug report.
+ LF bug 1731: wrong location of libraries for OpenBSD 64Bit architectures
+ High: crmd: Prevent shutdown hangs by allowing the crmd to forget about pending actions for deleted resources
+ High: PE: Relax an assumption regarding clones that is not true when they are unmanaged
+ Medium: PE: Bug 1722 - By default, exhibit the old start-failures-are-fatal behaviour regardless of how resource-failure-stickines is set
+ Remove one unlink() of the temporary hostcache file.
+ hb_uuid: Remove superfluous sync().
+ STONITH: external/ssh: Disable StrictHostKeyChecking and PasswordAuthentication by default.
+
RA: mysql: Allow arbitrary commandline arguments for mysqld + CTS: patterns adjusted for exiting processes (LF 1725).
+ High: PE: Prevent an infinit pe/te loop when reprobing with resources in master mode
+ mgmt: Provide "Default" setting for crm configurations
+ mgmt: Provide complete pengine and crmd configurations with dynamic rendering
+ mgmt: Prevent haclient from falling into an error caused by blank resource metadata
+ mgmt: Add "meta_attributes" support to mgmt
+ mgmt: Resolve target_role problem for sub-resource
+ High: PE: Prevent use of NULL in crm_mon when date_expressions are used by ensuring that data_set->now is always set
+ PE: High: Remove an errant call to exit() than prevented an assert from being triggered
+ Low: Build: Remove unused LDADD entries (libpils.la and libapphb.la)
+ High: PE: Clone colocation fixes
+ Low: PE: Extra debug information
+ Low: PE: Avoid pointless copying of XML items (actions) during graph creation
+ Low: PE: Log the current cluster state at LOG_NOTICE instead of LOG_INFO
+ Medium: PE: Master internal ordering enhancements. Added stopped->start, stopped->promote.
+ Medium: PE: Make sure all slaves (not just re-allocated ones) have role=Slave
+ Medium: PE: Fix minor memory leak when stopping orphaned resources
+ Medium: CRM: Re-evaluate appropriate IPC message queue lengths and throtle IPC
_clients_ that hit them
+ High: PE: Missing header file change from previous commit (cs: 0c14cfe57dd9)
+ High: PE: Ensure that resources depending (by order) on a master are not promoted if no master is available
+ Low: PE: Update regression tests - some action numbers changed
+ Low: Tools: crm_resource - send errors to stderr and to regular logging
+ Medium: DTD: Fix comment regarding rsc_ordering constraints
+ High: PE: Fix manditory ordering with m/s resources
+ Low: PE: clones - Remove debug logging and simplify asserts in expand_list()
+ Low: crmd: Update the CIB with the node's version data when it becomes DC
+ mgmt: Prevent clone and master_slave from belonging to group
+ mgmt: Make several fields of rsc_location editable
+ mgmt: Rename "Places" to "Locations" to match with the CIB name
+ mgmt: Specify the default type of added item to correspond to the object that the cursor focused on
+ mgmt: Add boolean_op for location rules in mgmt
+
RA mysql, pgsql: use getent(1) instead of /etc/passwd + stonithd (LF 1714,1727): fix a race condition
+ stonithd (LF 1726): a hostlist may be empty
+ stonith/ibmrsa: allow ',' in the hostlist
+ RA: Xinetd: Fix stop/monitor to not fail if service isn't available yet
+ CTS: Fix a path reference to LSBDummy, so that CTS has a chance of working.
+ RA: Raid1: Allow the homehost setting to be specified.
+ Novell 329833: dotted-quad netmask notation broken in findif
+ lrm stonith plugin: fix exit code handling
+ stonith/external: improve logging
+ stonithd/lib: replace cookie generation with uuid
+ stonithd: code review and cleanup
+ LRM test: test xml of stonith agents through stonithd (included in the BSC)
+ stonithd: have -a really mean startup alone
+ Low: Stonith: external/rackpdu - Fix the metadata short descriptions
+ cl_msg: fix a typo
+ stonithd: library code review and cleanup
+ stonithd: increase the maxdispatchtime
+ Debian: heartbeat doesn't depend on libperl-dev
+ lrmd (LF 1715): increase the max dispatch time for lrmd
+ Corrected spelling/grammar errors in BasicSanityCheck
+ Medium: PE: Enact a saner default for rsc_order.score (s/0/INFINITY)
+ High: PE: More generic method of detecting colocation loops in native_merge_weights
+ Medium: Logging: Sanitize and centrally define the log facility used by various subsystems
+ Low: Build: Build the contrib directory last
+ Low: cts: Fix the log facility map
+ Low: CRM: Key of the already set debug_level global variable when setting crm_log_level
+ Low: CRM: Downgrade digest logging now that its under control
+ Low: cib: Logging - version message should not have been an error
+ Low: PE: Logging - Include the transition_magic key when logging changed parameters
+ Low: CTS: Add extra ignore pattern when the CCM is shot - do it properly this time
+ Medium: Core: Ensure heartbeat flushes all its logs before exiting
+ Low: TE: Pause for a moment before trying to reconnect to stonithd
+ Medium: cib: Improved transparency when in operating in degraded mode
+ Low: crmd: Remove redundant code
+ Low: CTS: Add extra ignore pattern when the CCM is shot
+ Low: CTS: Add BadNews pattern - parameters should never change
+ Low: Admin: crmadmin - Logging tweak
+ Low: PE: More efficient logging of orphan resources
+ Medium: PE: Bug 1710 - A resource's failcount was ignored on nodes with no operations for the resource
+ LRM test: various cleanups
+ Added a script for creating the configuration for a single node
+ Low: PE: Prevent users from trying to use on_fail=fence and stonith-enabled=false
+ Low: CIB: Allow notification to remote clients
+ Low: PE: Downgrade logging when we fallback to default option values
+ High: Core: Prevent logging buffer overflow - ensure the entity is NULL terminated
+ High: PE: Make sure only the correct number of cloned groups are started
+ Low: PE: Stop the master placement score from affecting the synapse's priority
+ High: PE: Include preferences of colocated resources when promoting masters
+ High: RA: Remove bashism from Filesystem OCF agent
+ Low: CTS: Limit the number of PE input files created
+ Low: DTD: Include helpful comment for rsc_order constraints
+ Medium: PE: Bug #1705 - Don't segfault when allocating empty groups
+ Low: Stonith: external/ssh - simplify the reset command and wait the same amount of time as the non-external variant
+ Medium: RA: Treat migrating (status 1) as running to avoid
+ Low: Tools: Allow pingd to change identity based on the name of its executable - workaround for Bug #1701
+ High: PE: Ensure manditory ordering constraints behave correctly
+ proctrack/LRM (LF 1697): really remove sources (performance/memory leak fix)
+ mgmtd: allow old clients to connect
+ Low: PE: Regression test for group colocation
+ Medium: PE: Simplify the allocation of group resources
+ High: PE: Optimize the merge_weights functionfor groups
+ High: PE: Make sure the merge_weights function is called for complex resources
+ High: PE: More accurate method of detecting colocation loops in native_merge_weights
+ High: PE: Implement a smarter algorithm for colocation
+ Low: PE: Fold child-loop functions back into the ones doing the actual work
+ Medium: PE: Expose rsc->children which should mean we can delete all the 'loop' functions for complex objects
+ LRM test: report findings if the tests failed
+ RA: Filesystem: VARLIB -> RSCTMP
+ RA: Filesystem: Fix path reference.
+ Move INITDIR to proper include.
+ RA: o2cb: Fix path broken in recent "cleanups".
+ LRM test: include parts of lrm regression testing in the BasicSanityCheck
+ raexecocf: include time.h
+ lrmd: log non-recurring operations
+ lrmd: wrong reference in one debug message
+ stonithd (LF 1615): get the list of hosts from an RA asynchronously
+ High: crmd: Tell the LRM to refresh the resource's default parameters
+ LRM: add capability to copy parameters from op to rsc
+ raexecocf: a fix for bug introduced in 2a787bbd73f6; longer sleep in get_resource_meta
+ Medium: Stonith: prevent superfluous logging
+ High: DTD: Remove stary character that made the DTD invalid
+ Low: Admin: Bug 1691 - crm_resource doesn't complain about empty strings as property values
+ Low: CRM: DTD - Explicitly specify a default for rsc_colocation.symmetrical
+ Low: CRM: DTD - Expand the list of accepted values for boolean fields
+ Low: RA: ManageVE - Fix status when VE files are not persistent
+ lrmd: performance 2: remove unnecessary memstat
+ lrmd: performance 1: replace linear with binary search for msg type.
+ Corrected a directory name in the '%pre' script in the spec file
+ RA apache: drop annoying messages if no configuration files are found
+ Changed the specfile so that it uses only the correct name for python/gtk package according to the distro
+ [LDIRECTORD] Remove supervise-ldirectord-config
+ specfile: converted %elseif chains to %endif %if
+ apache RA (LF 1692): move sourcing
/resource.d/heartbeat/.ocf-shellfuncs before var reference
+ detect fedora correctly
+ a number of small fixes to make fedora packages correctly
+ Put in an APCmastersnmp.c patch from Peter Kruse which makes the plugin work with new APC master switches
+ LRM (LF 1684): a meta-data operation fix introducing short sleep on zero reads
+ configure.in: fix for platforms with libc in /lib/tls
+ Updated arguments to crm_log_init to be consistent with new API.
+ LF bug # 1662 - massive heartbeat specfile update - to make it more usable
+ Medium: RA: EvmsSCC - Handle start failures caused by peers starting at the same time
+ Low: logd: Update the default config file to reflect the new default log facility
+ High: clplumbing: Use the correct environment variable when deciding to log via ha_logd
+ Low: Admin: crmadmin - exit with rc=0 when the command completed successfully
+ Medium: PE: Bug 1685 - Supply the correct default for rsc_order constraints
+ Low: PE: Fix a parse error in the attr8 regression test
+ High: TE: target_rc should default to 0 (LRM_OP_DONE) to detect monitor actions that fail the first time they are executed
+ High: PE: Bug 1682 - Prevent use-of-NULL when clones have no active children
+
[STONITH] Disable ipmilan by default +
[STONITH] ipmilan: fix potential segvfault in options parsing + [RA] patch to fix bashism in resources/OCF/pgsql
+ LRM (LF 1458): include timing information in the op data.
+ RA: eDir88: fix to allow meta-data to work in case NDS is not installed
+ Put in updates to the Informix script as supplied by Lars Forseth
+ An open call with the O_CREAT flag was missing the file mode.
+ High: crmd: Sanitize the maintenance of node_state entries
+ High: cib: Support the removal of multiple child objects - required for subsequent commit
+ Low: cts: 'Updating failcount for' is no longer worth tracking (action failures show up as other ERROR logs)
+ Low: TE: Improved logging of unconfirmed actions
+ Medium: TE: Dont modify the failcount for 'pending' action updates (rc=-1)
+ Medium: admin: Set the exit code to non-zero if crmadmin can't connect to the crmd
+ High: PE: Break graph loops involving stonith + regression test
+ Low: PE: Improvements to the graph and .dot creation code
+ Medium: crm: Bug 1680 - parse_xml() can fail to detect mismatching close tags
+ High: PE: all_stopped should only depend on native resources to avoid recursion
+ High: PE: The all_stopped action should only be part of the graph when the stop actions are runnable
+ High: PE: Only create
_one_ pseudo op per name (all_stopped/stonith_up)
+ Low: PE: Logging of changed resource definition
+ High: TE: Allow reconnection to stonithd if it is respawned
+ High: stonithd: Sanitize the login/logout functions and add the ability to reset the callback IPC channel
+
[STONITH] ipmilan: accept documented auth and priv values + [BUILD] don't remove tarbals in clean target
+ [BUILD] don't ignore errors on install
+ [RA] patch to fix bashism in heartbeat/lib/hb_setsite.in
+ [RA] patch to fix bashism in resources/OCF/pgsql
+ [RA] patch to fix bashism in resources/OCF/o2cb
+ [RA] patch to fix bashism in heartbeat/lib/hb_setweight.in
+ [RA] patch to fix bashism in heartbeat/lib/hb_setsite.in
+ [RA] patch to fix bashism in heartbeat/lib/hb_delnode.in
+ [RA] patch to fix bashism in heartbeat/lib/hb_addnode.in
+ [RA] patch to fix bashism in resources/OCF/eDir
+ [RA] patch to fix bashism in resources/OCF/SysInfo
+ [RA] patch to fix bashism in resources/OCF/Stateful
+ [RA] patch to fix bashism in heartbeat/lib/LRMBasicSanityCheck.in
+ hbagent: add partial configuration
+ tools: ocf-tester: add capability to test with lrmadmin/lrmd
+ Low: PE: Handle the 'fail' task in text2task()
+ Medium: PE: Dont check the parameter digest of orphan operations
+ Low: PE: ptest - dont execute the graph by default
+ High: PE: check_action_definition() - Ensure the meta attributes are also set so that the digest is calculated correctly (and consistently with the crmd)
+ High: CRM: filter_action_parameters() - Make a copy of interval/timeout in as they're about to be free'd
+ Low: PE: Update expected outputs with new version number
+ Low: PE: Warn when multiple nodes have the same uname
+ Low: cib: Refactor reading and writing of the CIB to disk
+ Medium: cib: Ensure fflush() and fsync() are always called when writing XML digests to disk
+ Medium: CRM: Ensure fflush() and fsync() are always called when writing XML to disk
+ High: cib: Allow the CIB to reconnect if the CCM is killed and respawned
+ High: cib: Handle master updates that fail DTD validation
+ High: TE: Allow the TE to reconnect if stonithd is killed and respawned
+ Low: cts: ResourceRecover - Match the interval to that used by crm_resource
+ Low: Stonith: New external plugin for handling 'APC Switched Rack PDU AP7952' devices
+ [DEBIAN] make heartbeat binNMU safe
+ LOW: cts: Remove unneeded argument for DC_IDLE pattern
+ Low: cib: Enhancements to logging of DTD errors
+ Low: cib: Rearrange so code to avoid excessive indenting
+ Low: PE: Correct spelling in log message
+ Low: crm: Let the caller decide how to log DTD validation failures
+ Low: crmd: Ensure our assumption that all resources have a class and type is true
+ Low: cts: Update the expected log patterns and remove dups
+ Low: cts: Resurrect the ComponentFail test
+ Low: cts: make cluster_stable() more informative
+ High: crmd: Simplify node_state maintenance
+ High: CIB: Improved CCM interaction
+ Low: CIB: Minor logging improvement
+ High: crmd: Improvements to shutdown sequence (particularly in response to failures)
+ CTS: Low: Ignore the correct BadNews message in ResourceRecover
+ CTS: Low: Tell NearQuorumPointTest to look for Pat:DC_IDLE before declaring success
+ CTS: Low: Optimize filtering of BadNews
+ Admin: Low: Bug 1603 - Allow CIB digest files to be verified
+ RA: apache - make status quieter
+ [RA] eDir88: include the stop option
+ OSDL bug 1666: in BSC, make sure temp rsc dir exists for RAs
+ Contrib: dopd - Fix usage of crm_log_init() by code that shouldn't be using it
+ Tools: ocf-tester - use the default value for OCF_ROOT if it exists
+ RA: IPaddr2 - Make the check for the modprobe/iptables utilities conditional on the IP being cloned
+ CRM: Update crm/cib feature sets and the set of tags/attributes used for feature set detection
+ crmd: Simplify the detection of active actions and resources at shutdown
+ PE: Use failcount to handle failed stops and starts
+ TE: Set failcount to INFINITY for resources that fail to start or stop
+ CRM: Remove debug code that should not have been committed
+ PE: Add regression test for previous commit
+ PE: Regression: Allow M/S resources to be promoted based solely on rsc_location constraints
+ PE: Fix up the tests now that compare_version() functions correctly (as of cs: 7d69ef94a258)
+ CRM: Fix compare_version() to actually work correctly on a regular basis
+ PE: Update testcases to include all_stopped (added in cs: 800c2fec24ee)
+ crmd: Bug 1655 - crmd can't exit when the PE or TE is killed from underneath it
+ Tools: Bug 1653 - Misc attrd/attrd_updater cleanups
+ Tools: Bug 1653 - Further changes to prevent use of NULL when no attribute is specified
+ CRM: Make logging setup consistent and do not log the command-line to stderr
+ RA: Delay (v1) - Remove extra characters from call to ra_execocf
+ Tools: Bug 1653 - attrd crashes when no attribute is specified
+ OCF: Provide the location of /sbin as used by some agents (HA_SBIN_DIR)
+ PE: Move the creation of stonith shutdown constraints to native_internal_constraints()
+ crmd: Only remap monitor operation status to LRM_OP_DONE under the correct conditions
+ PE: Handle two new actions in text2task
+ CTS: Give stonith devices penty of time to start
+ PE: Include description for the remove-after-stop option
+ PE: Streamline STONITH ordering. Make sure 'all_stopped' depends on all STONITH ops.
+ PE: Aggregate the startup of fencing resources into a stonith_up pseudo-action
+ PE: STONITH Shutdown ordering
+ Bugzilla 1657: Speed up BasicSanityCheck and also make logging inheritance more uniform.
+ OSDL 1449 / Novell 291749: GUI should not overwrite more specific settings of contained resources.
+ Remove autoconf and friends on make distclean
------------------
PS: Special thanks to Dejan for making up this change log - it's an annoying and thankless task.
评论