1 %define initdir %{_sysconfdir}/init.d
5 Packager: Samba Team <samba@samba.org>
10 License: GNU GPL version 3
11 Group: System Environment/Daemons
12 URL: http://ctdb.samba.org/
14 Source: ctdb-%{version}.tar.gz
16 Prereq: /bin/mktemp /usr/bin/killall
17 Prereq: fileutils sed /etc/init.d
19 Provides: ctdb = %{version}
22 BuildRoot: %{_tmppath}/%{name}-%{version}-root
25 ctdb is the clustered database used by samba
28 #######################################################################
32 # setup the init script and sysconfig file
33 %setup -T -D -n ctdb-%{version} -q
38 if ccache -h >/dev/null 2>&1 ; then
46 ## always run autogen.sh
49 CFLAGS="$RPM_OPT_FLAGS $EXTRA -O0 -D_GNU_SOURCE -DCTDB_VERS=\"%{version}-%{release}\"" ./configure \
51 --sysconfdir=%{_sysconfdir} \
53 --localstatedir="/var"
55 make docdir=%{_docdir} showflags
56 make docdir=%{_docdir}
59 # Clean up in case there is trash left from a previous build
60 rm -rf $RPM_BUILD_ROOT
62 # Create the target build directory hierarchy
63 mkdir -p $RPM_BUILD_ROOT%{_sysconfdir}/sysconfig
64 mkdir -p $RPM_BUILD_ROOT%{_sysconfdir}/init.d
66 make DESTDIR=$RPM_BUILD_ROOT docdir=%{_docdir} install
68 install -m644 config/ctdb.sysconfig $RPM_BUILD_ROOT%{_sysconfdir}/sysconfig/ctdb
69 install -m755 config/ctdb.init $RPM_BUILD_ROOT%{initdir}/ctdb
71 mkdir -p $RPM_BUILD_ROOT%{_docdir}/ctdb/tests/bin
72 install -m755 tests/bin/ctdb_transaction $RPM_BUILD_ROOT%{_docdir}/ctdb/tests/bin
74 # Remove "*.old" files
75 find $RPM_BUILD_ROOT -name "*.old" -exec rm -f {} \;
78 rm -rf $RPM_BUILD_ROOT
81 #######################################################################
83 #######################################################################
88 %config(noreplace) %{_sysconfdir}/sysconfig/ctdb
89 %config(noreplace) %{_sysconfdir}/ctdb/notify.sh
90 %config(noreplace) %{_sysconfdir}/ctdb/ctdb-crash-cleanup.sh
91 %config(noreplace) %{_sysconfdir}/ctdb/functions
92 %attr(755,root,root) %{initdir}/ctdb
94 %{_docdir}/ctdb/README.eventscripts
95 %{_docdir}/ctdb/recovery-process.txt
96 %{_sysconfdir}/ctdb/events.d/00.ctdb
97 %{_sysconfdir}/ctdb/events.d/01.reclock
98 %{_sysconfdir}/ctdb/events.d/10.interface
99 %{_sysconfdir}/ctdb/events.d/11.natgw
100 %{_sysconfdir}/ctdb/events.d/11.routing
101 %{_sysconfdir}/ctdb/events.d/20.multipathd
102 %{_sysconfdir}/ctdb/events.d/31.clamd
103 %{_sysconfdir}/ctdb/events.d/40.vsftpd
104 %{_sysconfdir}/ctdb/events.d/41.httpd
105 %{_sysconfdir}/ctdb/events.d/50.samba
106 %{_sysconfdir}/ctdb/events.d/60.nfs
107 %{_sysconfdir}/ctdb/events.d/61.nfstickle
108 %{_sysconfdir}/ctdb/events.d/62.cnfs
109 %{_sysconfdir}/ctdb/events.d/70.iscsi
110 %{_sysconfdir}/ctdb/events.d/91.lvs
111 %{_sysconfdir}/ctdb/statd-callout
116 %{_bindir}/ctdb_diagnostics
118 %{_mandir}/man1/ctdb.1.gz
119 %{_mandir}/man1/ctdbd.1.gz
120 %{_mandir}/man1/onnode.1.gz
121 %{_includedir}/ctdb.h
122 %{_includedir}/ctdb_private.h
123 %{_libdir}/pkgconfig/ctdb.pc
125 %{_docdir}/ctdb/tests/bin/ctdb_transaction
128 * Tue Aug 24 2010 : Version 1.0.112-34
129 - Merge NFS auto-restart patch from master
130 * Tue Aug 11 2010 : Version 1.0.112-33
132 Add a new event "ipreallocated" so that we can update the natgw
133 configuration also when normal/light failovers occur and not just when
134 heavy/full database recoveries/cluster rebuilds trigger.
135 * Tue Aug 10 2010 : Version 1.0.112-32
136 - Dont check if natgw is enabled or not in the natgw script
137 - disable the check if winbind is ok
138 - optimize 61.nfstickles
139 - fix a fd leak from the abort-vacuuming patch
140 * Tue Aug 3 2010 : Version 1.0.112-31
141 - Retry the vacuum abort code.
143 - When rebalancing ip addresses, increase number of iterations from 5
144 - When using deterministic ips, dont force ips without home address
146 * Mon Jul 26 2010 : Version 1.0.112-30
147 - last minute fix for crash bug in the recent "vacuum abort" code.
149 * Mon Jul 26 2010 : Version 1.0.112-29
150 - Fix for a SEGV that can happen when tcp tickles fail/timeout
153 * Wed Jul 21 2010 : Version 1.0.112-28
154 - Turn back on deterministic IPs by default
155 (needs more analysis)
157 * Wed Jul 21 2010 : Version 1.0.112-27
158 - Turn off deterministic IPs by default
160 - vaccuum abort during freeze
161 CQ:S1018154 & S1018349
162 - logging enhancement for ctdbd children
164 * Thu Jul 15 2010 : Version 1.0.112-26
165 - iptables parallel invocation fix
168 * Tue Jul 6 2010 : Version 1.0.112-25
172 * Mon Jul 5 2010 : Version 1.0.112-24
173 - Extra logging on tdb_chainunlock failures.
174 - Extra sanity check in ctdb_become_dmaster
176 - More robustness against IDR wrap
178 - Recovery failure under stress fix
181 * Tue Jun 8 2010 : Version 1.0.112-23
182 - Fix a SEGV that can be triggered by "ctdb delip"
184 - Add iptables filters to stop clients from connecting to the NATGW
187 - Add timestamps to the ctdb statistics output
188 - Change "ctdb addip" to block until the address is active
190 - Add additional log messages when tdbs can no longer be locked/chain
193 * Mon May 24 2010 : Version 1.0.112-22
194 - Fix bug in 62.cnfs to allow exports that are quoted.
195 - Add monitoring og Quorum for the 62.cnfs script
196 - Allow restoredb to restore a backup to to a different database
197 * Fri May 21 2010 : Version 1.0.112-21
198 - Fix a bug where we would fail to remove the natgw configuration
199 and would leave a ip address for natgw still remaining on loopback.
201 - In the ctdb command line tool, we do not wait long enough for the ipreallocate
202 command to finish on the recovery daemon. which sometimes would timeout if the
203 recovery daemon was busy, or we had rolling recoveries.
205 - During failed recoveries we could get in a state where recovery attempts stopped
206 while database priority level 2 or 3 remain frozen.
208 * Tue May 11 2010 : Version 1.0.112-20
209 - Add number of recoveries done to the ctdb statistics output.
210 - Use the fuill range for IDR values.
212 - When performing a recovery, make sure all nodes agree with recmaster
215 - Lower the loglevel for some debug messages
217 * Tue May 4 2010 : Version 1.0.112-19
218 - Add an eventscript 62.cnfs to allow better integration with GPFS.
220 * Mon May 3 2010 : Version 1.0.112-18
221 - Make the NATGW eventscript check for when the system is misconfigured
222 to use the same address for both natgw as well as a public address
224 - A recent change to monitor the ip address assignment broke "ctdb moveip".
225 Fix ctdb moveip so it works again. BZ62782
226 * Fri Apr 23 2010 : Version 1.0.112-17
227 - In ctdb-crash-cleanup.sh also check for and remove the NATGW address
229 - Add monitoring and warning of low memory condition
230 CTDB_MONITOR_FREE_MEMORY_WARN
231 * Thu Apr 22 2010 : Version 1.0.112-16
232 - Fix for a syntax error bug in 10.interfaces
233 - Add a ctdb-crash-cleanup-sh script that can be used
234 from a cronjob to release ip addresses if ctdb has stopped.
235 * Wed Apr 21 2010 : Version 1.0.112-15
236 - Change how we add/remove iptable rules during recovery to make
237 rules leaks less likey.
238 - change a debug message loglevel from ERR to NOTICE
240 - In the recovery daemon, track pulbic ip assignment
241 across the cluster and verify consistency.
242 - From Martins: change the 10.interface script to handle
243 virtio interfaces correctly for virtual clusters.
244 - Make the recovery master verify the reclock setting across the cluster
245 and ban nodes with inconsistencies.
247 * Mon Mar 29 2010 : Version 1.0.112-14
248 - Wipe the new serverid database before starting winbind
249 * Wed Mar 25 2010 : Version 1.0.112-13
250 - Try to restart NFS if "service nfs start" failed.
252 * Wed Mar 11 2010 : Version 1.0.112-12
253 - From christian ambach : drop the loglevel of a vacuuming message
254 - From Wolfgang Mueller-Friedt : fix bug in ctdb_setstatus
255 - Use "service nfs status" instead of rpcinfo when probing nfs
256 - make sure to always create the tickles directory
257 * Thu Feb 25 2010 : Version 1.0.112-11
258 - Increase default script timeout to 90 seconds.
260 - Add a new tunable EventScriptLogTimeout which will log a message everytime
261 a monitor script takes longer than this to finish.
263 * Tue Feb 23 2010 : Version 1.0.112-10
264 - revert the change in 10.0.0.112-9 and make a new attempt to make the scripts behave.
265 - make writing the ticklelist in 61.nfstickle a background task to avoid
266 having a long cluster fs pause cause a node to become unhealthy
267 - critical bugfix. during an error path in the "end recovery" code
268 we could release a memory block before we had finished referencing it
269 which could lead to a segv. bz 61068
270 - make sure we tear down the natgw configuration when a node become stopped
271 or else we might end up with a duplicate ip address when a different node
272 takes over the natgw role. bz 61036
273 * Tue Feb 16 2010 : Version 1.0.112-9
274 - treat all scripts that hang (except for startup) as success even if they hung
275 * Mon Feb 15 2010 : Version 1.0.112-8
276 - dont increase the tdb seqnum when we just do a dmsater migration
277 this prevents trashing for notify.tdb and registry.tdb since
278 smbd uses tdb seqnums for these databases.
279 - try to restart rpc.rquotad if it is not running
280 * Thu Feb 11 2010 : Version 1.0.112-7
281 - Add logging of process if we detect we have started to swap from
282 the 00.ctdb eventscript.
283 - Reduce the log level of various messages to stop them spamming the log file
284 - Add a variable to disable checking if kNFSd has hung/died or not
285 * Tue Feb 9 2010 : Version 1.0.112-6
286 - Reduce the log level for debug messages related to tracking filedescriptors
288 - Update the linked list macros to provide efficient O(1) behaviour for
289 adding new entries to the tail of the list. This allows much larger lists.
290 - Increase the packet queue list to host 1000000 entries instead of 1000.
291 - when a smbd has hung and the packet queue has filled up. Close that client
292 instead of starting to discard packets. Discarding packets could discard
293 traversal related packets and could cause smbd to block indefinitely.
294 - fix the printout of high latency in the log.
295 * Tue Feb 2 2010 : Version 1.0.112-5
296 - Patch from Volker to fix a TDB transaction rolling back issue.
297 * Wed Jan 21 2010 : Version 1.0.112-4
298 - Update onnode with more flexible ways to define the path to the nodes file
299 * Wed Jan 20 2010 : Version 1.0.112-3
300 - Explicitely source the nfs sysconfig file from the 61.nfstickles script.
301 * Wed Jan 13 2010 : Version 1.0.112-2
302 - Incorrect use of dup2() could cause ctdb to spin eating 100% cpu.
303 * Tue Jan 12 2010 : Version 1.0.112
304 - Revert the use of wbinfo --ping-dc as it is proving too unreliable.
305 - Minor testsuite changes.
306 * Fri Dec 18 2009 : Version 1.0.111
307 - Fix a logging bug when an eventscript is aborted that could cause a crash.
308 - Add back cb_status that was lost in a previous commit.
309 * Fri Dec 18 2009 : Version 1.0.110
310 - Metxe: fix for filedescriptor leak in the new eventscript code.
311 - Rusty: fix for a crash bug in the eventscript code.
312 * Thu Dec 17 2009 : Version 1.0.109
313 - Massive eventscript updates. (bz58828)
314 - Nice the daemon instead of using realtime scheduler, also use mlockall() to
315 reduce the risk of blockign due to paging.
316 - Workarounds for valgrind when forking once for each script. Valgrind consumes
317 massive cpu when terminating the scripts on virtual systems.
318 - Sync the tdb library with upstream, and use the new TDB_DISALLOW_NESTING flag.
319 - Add new command "ctdb dumpdbbackup"
320 - Start using the new tdb check framework to validate tdb files upon startup.
321 - A new framework where we can control health for individual tdb databases.
322 - Fix a crash bug in the logging code.
323 - New transaction code for persistent databases.
324 - Various other smaller fixes.
325 * Mon Dec 7 2009 : Version 1.0.108
326 - Transaction updates from Michael Adam.
327 - Use the new wbinfo --ping-dc instead of -p in the eventscript for samba
328 the check if winbindd is ok.
329 - Add a better "process-exist" for samba so it will automatically
330 reap smbd's on stopped and banned nodes to reclaim subrecords.
331 This will be done a bit differently in the next release.
332 - Use a statically allocated buffer for the 'first-time' capture buffer
333 to reduce the pressure on malloc/free.
334 * Wed Dec 2 2009 : Version 1.0.107
335 - fix for rusty to solve a double-free that can happen when there are
336 multiple packets queued and the connection is destroyed before
337 all packets are processed.
338 * Tue Dec 1 2009 : Version 1.0.106
339 - Buildscript changes from Michael Adam
340 - Dont do a full recovery when there is a mismatch detected for ip addresses,
341 just do a less disruptive ip-reallocation
342 - When starting ctdbd, wait until all initial recoveries have finished
343 before we issue the "startup" event.
344 So dont start services or monitoring until the cluster has
346 - Major eventscript overhaul by Ronnie, Rusty and Martins and fixes of a few
348 * Thu Nov 19 2009 : Version 1.0.105
349 - Fix a bug where we could SEGV if multiple concurrent "ctdb eventscript ..."
350 are used and some of them block.
351 - Monitor the daemon from the syslog child process so we shutdown cleanly when
352 the main daemon terminates.
353 - Add a 500k line ringbuffer in memory where all log messages are stored.
354 - Add a "ctdb getlog <level>" command to pull log messages from the in memory
356 - From martin : fixes to cifs and nfs autotests
357 - from michael a : fix a bashism in 11.natgw
358 * Fri Nov 6 2009 : Version 1.0.104
359 - Suggestion from Metze, we can now use killtcp to kill local connections
360 for nfs so change the killtcp script to kill both directions of an NFS
362 We used to deliberately only kill one direction in these cases due to
364 - Suggestion from christian Ambach, when using natgw, try to avoid using a
365 UNHEALTHY node as the natgw master.
366 - From Michael Adam: Fix a SEGV bug in the recent change to the eventscripts
367 to allow the timeout to apply to each individual script.
368 - fix a talloc bug in teh vacuuming code that produced nasty valgrind
370 - From Rusty: Set up ulimit to create core files for ctdb, and spawned
371 processes by default. This is useful for debugging and testing but can be
372 disabled by setting CTDB_SUPRESS_COREFILE=yes in the sysconfig file.
373 - Remove the wbinfo -t check from the startup check that winbindd is happy.
374 - Enhance the test for bond devices so we also check if the sysadmin have
375 disabled all slave devices using "ifdown".
376 * Tue Nov 3 2009 : Version 1.0.103
377 - Dont use vacuuming on persistent databases
378 - Michael A : transaction updates to persistent databases
379 - Dont activate service automatically when installing the RPM. Leave this to the admin.
380 - Create a child process to send all log messages to, to prevent a hung/slow syslogd
381 from blocking the main daemon. In this case, discard log messages instead and let the child
383 - Michael A: updates to log messages
384 * Thu Oct 29 2009 : Version 1.0.102
385 - Wolfgang: fix for the vacuuming code
386 - Wolfgang: stronger tests for persistent database filename tests
387 - Improve the log message when we refuse to startup since wbinfo -t fails
388 to make it easier to spot in the log.
389 - Update the uptime command output and the man page to indicate that
390 "time since last ..." if from either the last recovery OR the last failover
391 - Michael A: transaction updates
392 * Wed Oct 28 2009 : Version 1.0.101
393 - create a separate context for non-monitoring events so they dont interfere with the monitor event
394 - make sure to return status 0 in teh callback when we abort an event
395 * Wed Oct 28 2009 : Version 1.0.100
396 - Change eventscript handling to allow EventScriptTimeout for each individual script instead of for all scripts as a whole.
397 - Enhanced logging from the eventscripts, log the name and the duration for each script as it finishes.
398 - Add a check to use wbinfo -t for the startup event of samba
399 - TEMP: allow clients to attach to databases even when teh node is in recovery mode
400 - dont run the monitor event as frequently after an event has failed
401 - DEBUG: in the eventloops, check the local time and warn if the time changes backward or rapidly forward
402 - From Metze, fix a bug where recovery master becoming unhealthy did not trigger an ip failover.
403 - Disable the multipath script by default
404 - Automatically re-activate the reclock checking if the reclock file is specified at runtime. Update manpage to reflect this.
405 - Add a mechanism where samba can register a SRVID and if samba unexpectedly disconnects, a message will be broadcasted to all other samba daemons.
406 - Log the pstree on hung scripts to a file in /tmp isntead of /var/log/messages
407 - change ban count before unhealthy/banned to 10
408 * Thu Oct 22 2009 : Version 1.0.99
409 - Fix a SEGV in the new db priority code.
410 - From Wolfgang : eliminate a ctdb_fatal() if there is a dmaster violation detected.
411 - During testing we often add/delete eventscripts at runtime. This could cause an eventscript to fail and mark the node unhealthy if an eventscript was deleted while we were listing the names. Handle the errorcode and make sure the node does not becomne unhealthy in this case.
412 - Lower the debuglevel for the messages when ctdb creates a filedescruiptor so we dont spam the logs with these messages.
413 - Dont have the RPM automatically restart ctdb
414 - Volker : add a missing transaction_cancel() in the handling of persistent databases
415 - Treat interfaces with the anme ethX* as bond devices in 10.interfaces so we do the correct test for if they are up or not.
416 * Tue Oct 20 2009 : Version 1.0.98
417 - Fix for the vacuuming database from Wolfgang M
418 - Create a directory where the test framework can put temporary overrides
419 to variables and functions.
420 - Wait a lot longer before shutting down the node when the reclock file
421 is incorrectly configured, and log where it is configured.
422 - Try to avoid running the "monitor" event when databases are frozen.
423 - Add logging for every time we create a filedescriptor so we can trap
425 * Thu Oct 14 2009 : Version 1.0.97
426 - From martins : update onnode.
427 Update onnode to allow specifying an alternative nodes file from
428 the command line and also to be able to specify hostnames on the
430 onnode host1,host2,...
431 * Tue Oct 13 2009 : Version 1.0.96
432 - Add more debugging output when eventscripts have trouble. Print a
433 "pstree -p" to the log when scripts have hung.
434 - Update the initscript, only print the "No reclock file used" warning
435 when we do "service ctdb start", dont also print them for all other
437 - When changing between unhealthy/healthy state, push a request to the
438 recovery master to perform an ip reallocation instead of waiting for the
439 recovery master to pull and check the state change.
440 - Fix a bug in the new db-priority handling where a pre-.95 recovery master
441 could no longer lock the databases on a post-.95 daemon.
442 - Always create the nfs state directories during the "monitor" event.
443 This makes it easier to configure and enable nfs at runtime.
444 - From Volker, forward-port a simper deadlock avoiding patch from the 1.0.82
445 branch. This is a simpler versionof the "db priority lock order" patch
446 that went into 1.0.95, and will be kept for a few versions until samba
447 has been updated to use the functionality from 1.0.95.
448 * Mon Oct 12 2009 : Version 1.0.95
449 - Add database priorities. Allow samba to set the priority of databases
450 and lock the databases in priority order during recovery
451 to avoid a deadlock when samba locks one database then blocks indefinitely
452 while waiting for the second databaso to become locked.
453 - Be aggressive and ban nodes where the recovery transaction start call
455 * Thu Oct 10 2009 : Version 1.0.94
456 - Be very aggressive and quickly ban nodes that can not freeze their databases
457 * Tue Oct 8 2009 : Version 1.0.93
458 - When adding an ip, make sure to update this assignment on all nodes
459 so it wont show up as -1 on other nodes.
460 - When adding an ip and immediately deleting it, it was possible that
461 the daemon would crash accessing already freed memory.
462 Readjust the memory hierarchy so the destructors are called in the right order.
463 - Add a handshake to the recovery daemon to eliminate some rare cases where
464 addip/delip might cause a recovery to occur.
465 - updated onnode documenation from Martin S
466 - Updates to the natgw eventscript to allow disabling natgw at runtime
467 * Fri Oct 2 2009 : Version 1.0.92
468 - Test updates and merge from martin
469 - Add notification for "startup"
470 - Add documentation for notification
471 - from martin, a fix for restarting vsftpd in the eventscript
472 * Tue Sep 29 2009 : Version 1.0.91
473 - New vacuum and repack design from Wolgang Mueller.
474 - Add a new eventscript 01.reclock that will first mark a node unhealthy and later ban the node if the reclock file can not be accessed.
475 - Add machinereadable output to the ctdb getreclock command
476 - merge transaction updates from Michael Adam
477 - In the new banning code, reset the culprit count to 0 for all nodes that could successfully compelte a full recovery.
478 - dont mark the recovery master as a ban culprit because a node in the cluster needs a recovery. this happens naturally when using ctdb recover command so dont make this cause a node to be banned.
479 * Sat Sep 12 2009 : Version 1.0.90
480 - Be more forgiving for eventscripts that hang during startup
481 - Fix for a banning bug in the new banning logic
482 * Thu Sep 3 2009 : Version 1.0.89
483 - Make it possible to manage winbind independently of samba.
484 - Add new prototype banning code
485 - Overwrite the vsftpd state file instead of appending. This eliminates
486 annoying errors in the log.
487 - Redirect some iptables commands to dev null
488 - From Michael A, explicitely set the broadcast when we takeover a public ip
489 - Remove a reclock file check we no longer need
490 - Skip any persistent database files ending in .bak
491 * Mon Aug 17 2009 : Version 1.0.88
492 - Add a new state for eventscripts : DISABLED.
493 Add two new commands "ctdb enablescript/disablescript" to enable/disable
494 eventscripts at runtime.
495 - Bugfixes for TDB from rusty.
496 - Merge/Port changes from upstream TDB library by rusty.
497 - Additional new tests from MartinS. Tests for stop/continue.
498 - Initial patch to rework vacuuming/repacking process from Wolfgang Mueller.
499 - Updates from Michael Adam for persistent writes.
500 - Updates from MartinS to handle the new STOPPED bit in the test framework.
501 - Make it possible to enable/disable the RECMASTER and LMASTER roles
502 at runtime. Add two new commands
503 "ctdb setlmasterrole/setrecmasterrole on/off"
504 - Make it possible to enable/disable the natgw feature at runtime. Add
505 the command "ctdb setnatgwstate on/off"
506 * Fri Jul 17 2009 : Version 1.0.87
507 - Add a new event "stopped" that is called when a node is stopped.
508 - Documentation of the STOPPED flag and the stop/continue commands
509 - Make it possible to start a node in STOPPED mode.
510 - Add a new node flag : STOPPED and commands "ctdb stop" "ctdb continue"
511 These commands are similar to "diasble/enable" but will also remove the node from the vnnmap, while disable only fails all ip addresses over.
512 - tests for NFS , CIFS by martins
513 - major updates to the init script by martins
514 - Send gratious arps with a 1.1 second stride instead of a 1 second stride to workaround interesting "features" of common linux stacks.
515 - Various test enhancements from martins:
516 - additional other tests
517 - add tests for grat arp generation, ping during failover, ssh and failover
518 - New/updated tcp tickle tests and supprot functions
519 - provide better debugging when a test fails
520 - make ctdbd restarts more reliable in the tests
521 - update the "wait bar" to make the wait progress in tests more obvious
523 - when dispatching a message to a handler, make the message a real talloc object so that we can reparent the object in the tallic hierarchy.
524 - document the ipreallocate command
525 - Updates to enable/disable to use the ipreallocate command to block until the following ipreallocation has completed.
526 - Update the main daemon and the tools to allow debug level to be a string instead of an integer.
527 - Update the sysconfig file to show using string literals instead of numeric values for the debuglevels used.
528 - If no debuglevel is specific, make "ctdb setdebug" show the available options.
529 - When trying to allocate network packets, add explicit checks if the network transport has been shutdown before trying and failing, to make log messages easier to read. Add this extra check and logging to every plave packets are allocated.
530 * Tue Jun 30 2009 : Version 1.0.86
531 - Do not access the reclock at all if VerifyRecoveryLock is zero, not even try to probe it.
532 - Allow setting the reclock file as "", which means that no reclock file at all should be used.
533 - Document that a reclock file is no longer required, but that it is dangerous.
534 - Add a control that can be used to set/clear/change the reclock file in the daemon during runtime.
535 - Update the recovery daemon to poll whether a reclock file should be sued and if so which file at runtime in each monitoring cycle.
536 - Automatically disable VerifyRecoveryLock everytime a user changes the location of the reclock file.
537 - do not allow the VerifyRecoveryLock to be set using ctdb setvar if there is no recovery lock file specified.
538 - Add two commands "ctdb getreclock" and "ctdb setreclock" to modify the reclock file.
539 * Tue Jun 23 2009 : Version 1.0.85
540 - From William Jojo : Dont use getopt on AIX
541 - Make it possible to use "ctdb listnodes" also when the daemon is not running
542 - Provide machinereadable output to "ctdb listnodes"
543 - Dont list DELETED nodes in the ctdb listnodes output
544 - Try to avoid causing a recovery for the average case when adding/deleting/moving an ip
545 - When banning a node, drop the IPs on that node only and not all nodes.
546 - Add tests for NFS and CIFS tickles
547 - Rename 99.routing to 11.routing so it executes before NFS and LVS scripts
548 - Increase the default timeout before we deem an unresponsive recovery daemon hung and shutdown
549 - Reduce the reclock timout to 5 seconds
550 - Spawn a child process in the recovery daemon ot check the reclock file to
551 avoid blocking the process if the underlying filesystem is unresponsive
552 - fix for filedescriptor leak when a child process timesout
553 - Dont log errors if waitpid() returns -1
554 - Onnode updates by Martins
555 - Test and initscript cleanups from Martin S
556 * Tue Jun 2 2009 : Version 1.0.84
557 - Fix a bug in onnode that could not handle dead nodes
558 * Tue Jun 2 2009 : Version 1.0.83
559 - Document how to remove a ndoe from a running cluster.
560 - Hide all deleted nodes from ctdb output.
561 - Lower the loglevel on some eventscript related items
562 - Dont queue packets to deleted nodes
563 - When building initial vnnmap, ignode any nonexisting nodes
564 - Add a new nodestate : DELETED that is used when deleting a node from an
566 - dont remove the ctdb socket when shutting down. This prevents a race in the
567 initscripts when restarting ctdb quickly after stopping it.
568 - TDB nesting reworked.
569 - Remove obsolete ipmux
570 - From Flavio Carmo Junior: Add eventscript and documentation for ClamAV antivirus engine
571 - From Sumit Bose: fix the regex in the test to handle the new ctdb
572 statistics output that was recently added.
573 - change the socket type we use for grauitious arps from the obsolete
574 AF_INET/SOCK_PACKET to instead use PF_PACKET/SOCK_RAW.
575 - Check return codes for some functions, from Sumit Bose, based on codereview by Jim Meyering.
576 - Sumit Bose: Remove structure memeber node_list_file that is no longer used.
577 - Sumit Bose: fix configure warning for netfilter.h
578 - Updates to the webpages by Volker.
579 - Remove error messages about missing /var/log/log.ctdb file from ctdb_diagnostics.sh from christian Ambach
580 - Additional error logs if hte eventscript switching from dameon to client mode fails.
581 - track how long it takes for ctdbd and the recovery daemon to perform the rec-lock fcntl() lock attemt and show this in the ctdb statistics output.
582 * Thu May 14 2009 : Version 1.0.82
583 - Update the "ctdb lvsmaster" command to return -1 on error.
584 - Add a -Y flag to "ctdb lvsmaster"
585 - RHEL5 apache leaks semaphores when crashing. Add semaphore cleanup to the
586 41.httpd eventscript and try to restart apache when it has crashed.
587 - Fixes to some tests
588 - Add a -o option to "onnode" which will redirect all stdout to a file for
590 - Add a natgw and a lvs node specifier to onnode so that we can use
592 - Assign the natgw address to lo instead of the private network so it can also
593 be used where private and public networks are the same.
594 - Add GPL boilerplates to two missing scripts.
595 - Change the natgw prefix NATGW_ to CTDB_NATGW_
596 * Fri May 8 2009 : Version 1.0.81
597 - use smbstatus -np instead of smbstatus -n in the 50.samba eventscript
598 since this avoids performing an expensive traverse on the locking and brlock
600 - make ctdb automatically terminate all traverse child processes clusterwide
601 associated to a client application that terminates before the traversal is
603 - From Sumit Bose : fixes to AC_INIT handling.
604 - From Michael Adam, add Tridge's "ping_pong" tool the the ctdb distro since
605 this is very useful for testing the backend filesystem.
606 - From Sumit bose, add support for additional 64 bit platforms.
607 - Add a link from the webpage to Michael Adams SambaXP paper on CTDB.
608 * Fri May 1 2009 : Version 1.0.80
609 - change init shutdown level to 01 for ctdb so it stops before any of the other services
610 - if we can not pull a database from a remote node during recovery, mark that node as a culprit so it becomes banned
611 - increase the loglevel when we volunteer to drop all ip addresses after beeing in recovery mode for too long. Make this timeout tuneable with "RecoveryDropAllIPs" and have it default to 60 seconds
612 - Add a new flag TDB_NO_NESTING to the tdb layer to prevent nested transactions which ctdb does not use and does not expect. Have ctdb set this flag to prevent nested transactions from occuring.
613 - dont unconditionally kill off ctdb and restrat it on "service ctdb start". Fail "service ctdb start" with an error if ctdb is already running.
614 - Add a new tunable "VerifyRecoveryLock" that can be set to 0 to prevent the main ctdb daemon to verify that the recovery master has locked the reclock file correctly before allowing it to set the recovery mode to active.
615 - fix a cosmetic bug with ctdb statistics where certain counters could become negative.
616 * Wed Apr 8 2009 : Version 1.0.79
617 - From Mathieu Parent: add a ctdb pkgconfig file
619 - add a funciton remove_ip to safely remove an ip from an interface, taking care to workaround an issue with linux alias interfaces.
620 - Update the natgw eventscript to use the safe remove_ip() function
621 - fix a bug in the eventscript child process that would cause the socket to be removed.
622 - dont verify nodemap on banned nodes during cluster monitoring
623 - Update the dodgy SeqnumInterval to have ms resolution
624 * Tue Mar 31 2009 : Version 1.0.78
625 - Add a notify mechanism so we can send snmptraps/email to external management systems when the node becomes unhealthy
626 - include 11.natgw eventscript in thew install so that the NATGW feature works
627 * Tue Mar 31 2009 : Version 1.0.77
628 - Update the 99.routing eventscript to also try to add the routes (back) during a releaseip event. Similar to the reasons why we must add addresses back during releaseip in 10.interfaces
629 * Wed Mar 24 2009 : Version 1.0.76
630 - Add a debugging command "xpnn" which can print the pnn of the node even when ctdbd is not running.
631 - Redo the NATGW implementation to allow multiple disjoing NATGW groups in the same cluster.
632 * Tue Mar 24 2009 : Version 1.0.75
633 - Various updates to LVS
634 - Fix a bug in the killtcp control where we did not set the port correctly
635 - add a new "ctdb scriptstatus" command that shows the status of the eventrscripts.
636 * Mon Mar 16 2009 : Version 1.0.74
637 - Fixes to AIX from C Cowan.
638 - Fixes to ctdb_diagnostics so we collect correct GPFS data
639 - Fixes to the net conf list command in ctdb_diagnostics
640 - Check the static-routes file IFF it exists in ctdb_diagnostics
641 * Wed Mar 4 2009 : Version 1.0.73
642 - Add possibility to disable the check of shares for NFS and Samba
643 - From Sumit Bose, fix dependencies so make -j works
644 * Wed Feb 18 2009 : Version 1.0.72
645 - Updates to test scripts by martin s
646 - Adding a COPYING file
647 - Use netstat to check for services and ports and fallback to netcat
648 only if netstat is unavailable.
649 * Mon Feb 2 2009 : Version 1.0.71
650 - Additional ipv6 fixes from Michael Adams
651 * Fri Jan 16 2009 : Version 1.0.70
652 - IPv6 support is completed. this is backward compatible with ipv4-only
653 systems. To use IPv6 with samba and ctdb you need current GIT of samba 3.3
654 or michael adams samba-ctdeb branch.
655 - Many enhancements to the build system and scripts to make it more SUSE
656 friendly by Michael Adams.
657 - Change of how the naming of the package is structured. We are now
658 using "1.0.70" as a release and "-1" as the revision instead of as
659 previously using "1.0" as release and ".70" as the revision.
661 * Thu Dec 18 2008 : Version 1.0.69
662 - Various fixes to scripts by M Adam
663 - Dont call ctdb_fatal() when the transport is down during shutdown
664 * Fri Dec 12 2008 : Version 1.0.68
665 - Fixes for monitoring of interfaces status from Michael Adam.
666 - Use -q instead of >/dev/null for grep to enhance readability of the
667 scripts from Michael Adam.
668 - Update to the "ctdb recover" command. This command now block until the
669 has completed. This makes it much easier to use in scripts and avoids
670 the common workaround :
672 ... loop while waiting for recovery completes ...
674 - Add a CTDB_TIMEOUT variable. If set, this variable provides an automatic
675 timeout for "ctdb <command>", similar to using -T <timeout>
676 - Set a unique errorcode for "ctdb <command>" when it terminates due to a
677 timeout so that scripts can distinguish between a hung command and what was
679 - Update "ctdb ban/unban" so that if the cluster is in recovery these commands
680 blocks and waits until after recovery is complete before the perform the
681 ban/unban operation. This is necessary since the recovery process can cause
682 nodes to become automatically unbanned.
683 - Update "ctdb ban/unban" to block until the recovery that will follow shortly
684 after this command has completed.
685 This makes it much easier to use in scripts and avoids the common
688 ... loop while waiting for recovery completes ...
690 - Bugfix for the new flags handling in 1.0.67. Abort and restart monitoring
691 if we failed to get proper nodemaps from a remote node instead of
692 dereferencing a null pointer.
693 - If ctdbd was explicitely started with the '--socket' argument, make
694 ctdbd automatically set CTDB_SOCKET to the specified argument.
695 This ensures that eventscripts spawned by the ctdb daemon will default to
696 using the same socket and talk to the correct daemon.
697 This primarily affects running multiple daemons on the same host and where
698 you want each instance of ctdb daemons have their eventscripts talk to the
700 - Update "ctdb ping" to return an error code if the ping fail so that it
701 can be used in scripts.
702 - Update to how to synchronize management of node flags across the cluster.
703 * Thu Dec 3 2008 : Version 1.0.67
704 - Add a document describing the recovery process.
705 - Fix a bug in "ctdb setdebug" where it would refuse to set a negative
707 - Print the list of literals for debug names if an invalid one was given
709 - Redesign how "ctdb reloadnodes" works and reduce the amont of tcp teardowns
710 used during this event.
711 - Make it possible to delete a public ip from all nodes at once using
713 * Mon Nov 24 2008 : Version 1.0.66
714 - Allow to change the recmaster even when we are not frozen.
715 - Remove two redundant SAMBA_CHECK variables from the sysconf example
716 - After a node failure it can take very long before some lock operations
717 ctdb needs to perform are allowed/works with gpfs again. Workaround this
718 by treating a hang/timeout as success.
719 - Dont override CTDB_BASE is fet in the shell already
720 - Always send keepalive packets regardless of whether the link is idle or not.
721 - Rewrite the disable/enable flag update logic to prevent a race between
722 "ctdb disable/enable" and the recovery daemon when updating the flags to
724 * Thu Nov 13 2008 : Version 1.0.65
725 - Update the sysconfig example: The default debug level is 2 (NOTICE) and not
727 - Add support for a CTDB_SOCKET environment variable for the ctdb command
728 line tool. If set, this overrides the default socket the ctdb tool will
730 - Add logging of high latency operations.
731 * Mon Oct 22 2008 : Version 1.0.64
732 - Add a context and a timed event so that once we have been in recovery for
733 too long we drop all public addresses.
734 * Mon Oct 20 2008 : Version 1.0.63
735 - Remove logging of "periodic cleanup ..." in 50.samba
736 - When we reload a nodes file, we must detect this and reload the file also
737 in the recovery daemon before we try to dereference somethoung beyond the end
739 * Thu Oct 16 2008 : Version 1.0.62
740 - Allow multiple eventscritps using the same prefix number.
741 It is undefined which order scripts with the same prefix will execute in.
742 * Wed Oct 15 2008 : Version 1.0.61
743 - Use "route add -net" instead of "ip route add" when adding routes in 99.routing
744 - lower the loglevel os several debug statements
745 - check the status returned from ctdb_ctrl_get_tickles() before we try to print them out to the screen.
746 - install a new eventscript 20.multipathd whoich can be used to monitor that multipath devices are healthy
747 * Wed Oct 15 2008 : Version 1.0.60
748 - Verify that nodes we try to ban/unban are reachable and print an error othervise.
749 - Update the client and server sides of TAKEIP/RELEASEIP/GETPUBLICIPS and GETNODEMAP to fall back to the old style ipv4-only controls if the new ipv4/ipv6 controls fail. This allows an ipv4/v6 enabled ctdb daemon to interoperate with earlier ipv4-only versions of the daemons.
750 - From Mathieu Parent : log debian systems log the package versions in ctdb diagnostics
751 - From Mathieu Parent : specify logdir location for debian (this patch was later reversed)
752 - From Michael Adams : allow # comments in nodes/public_addresses files
753 * Tue Oct 7 2008 : Version 1.0.59
754 - Updated "reloadnodes" logic. Instead of bouncing the entire tcp layer it is sufficient to just close and reopen all outgoing tcp connections.
755 - New eventscript 99.routing which can be used to re-attach routes to public interfaces after a takeip event. (routes may be deleted by the kernel when we release an ip)
756 - IDR tree fix from Jim Houston
757 - Better handling of critical events if the local clock is suddenly changed forward by a lot.
758 - Fix three slow memory leaks in the recovery daemon
759 - New ctdb command : ctdb recmaster which prints the pnn of the recmaster
760 - Onnode enhancements from Martin S : "healthy" and "connected" are now possible nodespecifiers
761 - From Martin S : doc fixes
762 - lowering some debug levels for some nonvital informational messages
763 - Make the daemon daemon monitoring stronger and allow ctdbd to detect a hung
765 - From C Cowan : patches to compile ipv6 under AIX
766 - zero out some structs to keep valgrind happy
767 * Wed Aug 27 2008 : Version 1.0.58
768 - revert the name change tcp_tcp_client back to tcp_control_tcp so
770 - Updates to the init script from Abhijith Das <adas@redhat.com>
771 * Mon Aug 25 2008 : Version 1.0.57
772 - initial support for IPv6
773 * Mon Aug 11 2008 : Version 1.0.56
774 - fix a memory leak in the recovery daemon.
775 * Mon Aug 11 2008 : Version 1.0.55
776 - Fix the releaseip message we seond to samba.
777 * Fri Aug 8 2008 : Version 1.0.54
778 - fix a looping error in the transaction code
779 - provide a more detailed error code for persistent store errors
780 so clients can make more intelligent choices on how to try to recover
781 * Thu Aug 7 2008 : Version 1.0.53
782 - Remove the reclock.pnn file it can cause gpfs to fail to umount
783 - New transaction code
784 * Mon Aug 4 2008 : Version 1.0.52
785 - Send an explicit gratious arp when starting sending the tcp tickles.
786 - When doing failover, issue a killtcp to non-NFS/non-CIFS clients
787 so that they fail quickly. NFS and CIFS already fail and recover
789 - Update the test scripts to handle CTRL-C to kill off the test.
790 * Mon Jul 28 2008 : Version 1.0.51
791 - Strip off the vlan tag from bond devices before we check in /proc
792 if the interface is up or not.
793 - Use testparm in the background in the scripts to allow probing
794 that the shares do exist.
795 - Fix a bug in the logging code to handle multiline entries better
796 - Rename private elements from private to private_data
797 * Fri Jul 18 2008 : Version 1.0.50
798 - Dont assume that just because we can establish a TCP connection
799 that we are actually talking to a functioning ctdb daemon.
800 So dont mark the node as CONNECTED just because the tcp handshake
802 - Dont try to set the recmaster to ourself during elections for those
803 cases we know this will fail. To remove some annoying benign but scary
804 looking entries from the log.
805 - Bugfix for eventsystem for signal handling that could cause a node to
807 * Thu Jul 17 2008 : Version 1.0.49
808 - Update the safe persistent update fix to work with unpatched samba
810 * Thu Jul 17 2008 : Version 1.0.48
811 - Update the spec file.
812 - Do not start new user-triggered eventscripts if we are already
813 inside recovery mode.
814 - Add two new controls to start/cancel a persistent update.
815 A client such as samba can use these to tell ctdbd that it will soon
816 be writing directly to the persistent database tdb file. So if
817 samba is -9ed before it has eitehr done the persistent_store or
818 canceled the operation, ctdb knows that the persistent databases
819 'may' be out of sync and therefore a full blown recovery is called for.
820 - Add two new options :
821 CTDB_SAMBA_SKIP_CONF_CHECK and CTDB_SAMBA_CHECK_PORTS that can be used
822 to override what checks to do when monitoring samba health.
823 We can no longer use the smbstatus, net or testparm commands to check
824 if samba or its config is healthy since these commands may block
825 indefinitely and thus can not be used in scripts.
826 * Fri Jul 11 2008 : Version 1.0.47
827 - Fix a double free bug where if a user striggered (ctdb eventscript)
828 hung and while the timeout handler was being processed a new user
829 triggered eventscript was started we would free state twice.
830 - Rewrite of onnode and associated documentation.
831 * Thu Jul 10 2008 : Version 1.0.46
832 - Document both the LVS:cingle-ip-address and the REMOTE-NODE:wan-accelerator
834 - Add commands "ctdb pnn", "ctdb lvs", "ctdb lvsmaster".
835 - LVS improvements. LVS is the single-ip-address mode for a ctdb cluster.
836 - Fixes to supress rpmlint warnings
838 - Change \s to [[:space:]] in some scripts. Not all RHEL5 packages come
839 with a egrep that handles \s even same version but different arch.
840 - Revert the change to NFS restart. CTDB should NOT attempt to restart
842 - Rewrite of the waitpid() patch to use the eventsystem for handling
844 * Tue Jul 8 2008 : Version 1.0.45
845 - Try to restart the nfs service if it has failed to respond 3 times in a row.
846 - waitpid() can block if the child does not respond promptly to SIGTERM.
847 ignore all SIGCHILD signals by setting SIGCHLD to SIG_DEF.
848 get rid of all calls to waitpid().
849 - make handling of eventscripts hanging more liberal.
850 only consider the script to have failed and making the node unhealthy
851 IF the eventscript terminated wiht an error
852 OR the eventscript hung 5 or more times in a row
853 * Mon Jul 7 2008 : Version 1.0.44
854 - Add a CTDB_VALGRIND option to /etc/sysconfig/ctdb to make it start
855 ctdb under valgrind. Logs go to /var/log/ctdb_valgrind.PID
856 - Add a hack to show the control opcode that caused uninitialized data
857 in the valgrind output by encoding the opcode as the line number.
858 - Initialize structures and allocated memory in various places in
859 ctdb to make it valgrind-clean and remove all valgrind errors/warnings.
860 - If/when we destroy a lockwait child, also make sure we cancel any pending transactions
861 - If a transaction_commit fails, delete/cancel any pending transactions and
862 return an error instead of calling ctdb_fatal()
863 - When running ctdb under valgrind, make sure we run it with --nosetsched and also
864 ensure that we do not use mem-mapped i/o when accessing the tdb's.
865 - zero out ctdb->freeze_handle when we free/destroy a freeze-child.
866 This prevent a heap corruption/ctdb crash bug that could trigger
867 if the freeze child times out.
868 - we dont need to explicitely thaw the databases from the recovery daemon
869 since this is done implicitely when we restore the recovery mode back to normal.
870 - track when we start and stop a recovery. Add the 'time it took to complete the
871 recovery' to the 'ctdb uptime' output.
872 Ensure by tracking the start/stop recovery timestamps that we do not
873 check that the ip allocation is consistend from inside the recovery daemon
874 while a different node (recovery master) is performing a recovery.
875 This prevent a race that could cause a full recovery to trigger if the
876 'ctdb disable/enable' commands took very long.
877 - The freeze child indicates to the master daemon that all databases are locked
878 by writing data to the pipe shared with the master daemon.
879 This write sometimes fail and thus the master daemon never notices that the databases
880 are locked cvausing long timeouts and extra recoveries.
881 Check that the write is successful and try the write again if it failed.
882 - In each node, verify that the recmaster have the right node flags for us
883 and force a push of our flags to the recmaster if wrong.
884 * Tue Jul 1 2008 : Version 1.0.43
885 - Updates and bugfixes to the specfile to keep rpmlint happy
886 - Force a global flags update after each recovery event.
887 - Verify that the recmaster agrees with our node flags and update the
889 - When writing back to the parent from a freeze-child across the pipe,
890 loop over the write in case the write failed with an error othervise
891 the parent will never be notified tha the child has completed the operation.
892 - Automatically thaw all databases when recmaster marks us as being in normal
893 mode instead of recovery mode.
894 * Fri Jun 13 2008 : Version 1.0.42
895 - When event scripts have hung/timedout more than EventScriptBanCount times
896 in a row the node will ban itself.
897 - Many updates to persistent write tests and the test scripts.
898 * Wed May 28 2008 : Version 1.0.41
899 - Reactivate the safe writes to persistent databases and solve the
900 locking issues. Locking issues are solved the only possible way,
901 by using a child process to do the writes. Expensive and slow but... .
902 * Tue May 27 2008 : Version 1.0.40
903 - Read the samba sysconfig file from the 50.samba eventscript
904 - Fix some emmory hierarchical bugs in the persistent write handling
905 * Thu May 22 2008 : Version 1.0.39
906 - Moved a CTDB_MANAGES_NFS, CTDB_MANAGES_ISCSI and CTDB_MANAGES_CSFTPD
907 into /etc/sysconfig/ctdb
908 - Lowered some debug messages to not fill the logfile with entries
909 that normally occur in the default configuration.
910 * Fri May 16 2008 : Version 1.0.38
911 - Add machine readable output support to "ctdb getmonmode"
912 - Lots of tweaks and enhancements if the event scripts are "slow"
913 - Merge from tridge: an attempt to break the chicken-and-egg deadlock that
914 net conf introduces if used from an eventscript.
915 - Enhance tickles so we can tickle an ipv6 connection.
916 - Start adding ipv6 support : create a new container to replace sockaddr_in.
917 - Add a checksum routine for ipv6/tcp
918 - When starting up ctdb, let the init script do a tdbdump on all
919 persistent databases and verify that they are good (i.e. not corrupted).
920 - Try to use "safe transactions" when writing to a persistent database
921 that was opened with the TDB_NOSYNC flag. If we can get the transaction
922 thats great, if we cant we have to write anyway since we cant block here.
923 * Mon May 12 2008 : Version 1.0.37
924 - When we shutdown ctdb we close the transport down before we run the
925 "shutdown" eventscripts. If ctdb decides to send a packet to a remote node
926 after we have shutdown the transport but before we have shutdown ctdbd
927 itself this could lead to a SEGV instead of a clean shutdown. Fix.
928 - When using the "exportfs" command to extract which NFS export directories
929 to monitor, exportfs violates the "principle of least surprise" and
930 sometimes report a single export line as two lines of text output
931 causing the monitoring to fail.
932 * Fri May 9 2008 : Version 1.0.36
933 - fix a memory corruption bug that could cause the recovery daemon to crash.
934 - fix a bug with distributing public ip addresses during recovery.
935 If the node that is the recovery master did NOT use public addresses,
936 then it assumed that no other node in the cluster used them either and
937 thus skipped the entire step of reallocating public addresses.
938 * Wed May 7 2008 : Version 1.0.35
939 - During recovery, when we define the new set of lmasters (vnnmap)
940 only consider those nodes that have the can-be-lmaster capability
941 when we create the vnnmap. unless there are no nodes available which
942 supports this capability in which case we allow the recmaster to
943 become lmaster capable (temporarily).
944 - Extend the async framework so that we can use paralell async calls
945 to controls that return data.
946 - If we do not have the "can be recmaster" capability, make sure we will
947 lose any recmaster elections, unless there are no nodes available that
948 have the capability, in which case we "take/win" the election anyway.
949 - Close and reopen the reclock pnn file at regular intervals.
950 Make it a non-fatal event if we occasionally fail to open/read/write
952 - Monitor that the recovery daemon is still running from the main ctdb
953 daemon and shutdown the main daemon when recovery daemon has terminated.
954 - Add a "ctdb getcapabilities" command to read the capabilities off a node.
955 - Define two new capabilities : can be recmaster and can be lmaster
956 and default both capabilities to YES.
957 - Log denied tcp connection attempts with DEBUG_ERR and not DEBUG_WARNING
958 * Thu Apr 24 2008 : Version 1.0.34
959 - When deleting a public ip from a node, try to migrate the ip to a different
961 - Change catdb to produce output similar to tdbdump
962 - When adding a new public ip address, if this ip does not exist yet in
963 the cluster, then grab the ip on the local node and activate it.
964 - When a node disagrees with the recmaster on WHO is the recmaster, then
965 mark that node as a recovery culprit so it will eventually become
967 - Make ctdb eventscript support the -n all argument.
968 * Thu Apr 10 2008 : Version 1.0.33
969 - Add facilities to include site local adaptations to the eventscript
970 by /etc/ctdb/rc.local which will be read by all eventscripts.
971 - Add a "ctdb version" command.
972 - Secure the domain socket with proper permissions from Chris Cowan
973 - Bugfixes for AIX from Chris Cowan
974 * Wed Apr 02 2008 : Version 1.0.32
975 - Add a control to have a node execute the eventscripts with arbitrary
976 command line arguments.
977 - Add a control "rddumpmemory" that will dump the talloc memory allocations
978 for the recovery daemon.
979 - Decorate the talloc memdump to produce better and easier memory leak
981 - Update the RHEL5 iscsi tgtd scripts to allow one iscsi target for each
983 - Add two new controls "addip/delip" that can be used to add/remove public
984 addresses to a node at runtime. After using these controls a "ctdb recover"
985 ir required to make the changes take.
986 - Fix a couple of slow memory leaks.
987 * Tue Mar 25 2008 : Version 1.0.31
988 - Add back controls to disable/enable monitoring on a node.
989 - Fix a memory leak where we used to attach CALL data to the ctdb structure
990 when performing a local call. Memory which would be lost if the call was
992 - Reduce the loglevel for the log output when someone connects to a non
993 public ip address for samba.
994 - Redo and optimize the vacuuming process to send only one control to each
995 other node containing all records to be vacuumed instead of one
996 control per node per record.
997 * Tue Mar 04 2008 : Version 1.0.30
998 - Update documentation cor new commands and tuneables
999 - Add machinereadable output to the ip,uptime and getdebug commands
1000 - Add a moveip command to manually failover/failback public ips
1001 - Add NoIPFallback tuneable that prevents ip address failback
1002 - Use file locking inside the CFS as alternative to verify when other nodes
1003 Are connected/disconnected to be able to recover from split network
1004 - Add DisableWhenUnhealthy tunable
1005 - Add CTDB_START_AS_DISABLED sysconfig param
1006 - Add --start-as-disabled flag to ctdb
1007 - Add ability to monitor for OOM condition
1008 * Thu Feb 21 2008 : Version 1.0.29
1009 - Add a new command to make expansion of an existing cluster easier
1010 - Fix bug with references to freed objects in the ctdb structure
1011 - Propagate debuglevel changes to the recovery daemon
1012 - Merge patches to event scripts from Mathieu Parent :
1013 - MP: Simulate "service" on systems which do not provide this tool
1014 - MP: Set correct permissions for events.d/README
1015 - Add nice helper functions to start/stop nfs from the event scripts
1016 * Fri Feb 08 2008 : Version 1.0.28
1017 - Fix a problem where we tried to use ethtool on non-ethernet interfaces
1018 - Warn if the ipvsadm packege is missing when LVS is used
1019 - Dont use absolute pathnames in some of the event scripts
1020 - Fix for persistent tdbs growing inifinitely.
1021 * Wed Feb 06 2008 : Version 1.0.27
1022 - Add eventscript for iscsi
1023 * Thu Jan 31 2008 : Version 1.0.26
1024 - Fix crashbug in tdb transaction code
1025 * Tue Jan 29 2008 : Version 1.0.25
1026 - added async recovery code
1027 - make event scripts more portable
1028 - fixed ctdb dumpmemory
1029 - more efficient tdb allocation code
1030 - improved machine readable ctdb status output
1032 * Wed Jan 16 2008 : Version 1.0.24
1033 - added syslog support
1034 - documentation updates
1035 * Wed Jan 16 2008 : Version 1.0.23
1036 - fixed a memory leak in the recoveryd
1037 - fixed a corruption bug in the new transaction code
1038 - fixed a case where an packet for a disconnected client could be processed
1039 - added http event script
1040 - updated documentation
1041 * Thu Jan 10 2008 : Version 1.0.22
1042 - auto-run vacuum and repack ops
1043 * Wed Jan 09 2008 : Version 1.0.21
1044 - added ctdb vacuum and ctdb repack code
1045 * Sun Jan 06 2008 : Version 1.0.20
1046 - new transaction based recovery code
1047 * Sat Jan 05 2008 : Version 1.0.19
1048 - fixed non-master bug
1049 - big speedup in recovery for large databases
1050 - lots of changes to improve tdb and ctdb for high churn databases
1051 * Thu Dec 27 2007 : Version 1.0.18
1052 - fixed crash bug in monitor_handler
1053 * Tue Dec 04 2007 : Version 1.0.17
1054 - fixed bugs related to ban/unban of nodes
1055 - fixed a race condition that could lead to monitoring being permanently disabled,
1056 which would lead to long recovery times
1057 - make deterministic IPs the default
1058 - fixed a bug related to continuous recovery
1059 - added a debugging option --node-ip