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: /sbin/chkconfig /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
39 ## always run autogen.sh
42 CFLAGS="$RPM_OPT_FLAGS $EXTRA -O0 -D_GNU_SOURCE -DCTDB_VERS=\"%{version}-%{release}\"" ./configure \
44 --sysconfdir=%{_sysconfdir} \
46 --localstatedir="/var"
52 # Clean up in case there is trash left from a previous build
53 rm -rf $RPM_BUILD_ROOT
55 # Create the target build directory hierarchy
56 mkdir -p $RPM_BUILD_ROOT%{_sysconfdir}/sysconfig
57 mkdir -p $RPM_BUILD_ROOT%{_sysconfdir}/init.d
59 make DESTDIR=$RPM_BUILD_ROOT install
61 install -m644 config/ctdb.sysconfig $RPM_BUILD_ROOT%{_sysconfdir}/sysconfig/ctdb
62 install -m755 config/ctdb.init $RPM_BUILD_ROOT%{initdir}/ctdb
64 # Remove "*.old" files
65 find $RPM_BUILD_ROOT -name "*.old" -exec rm -f {} \;
68 rm -rf $RPM_BUILD_ROOT
71 [ -x /sbin/chkconfig ] && /sbin/chkconfig --add ctdb
75 [ -x /sbin/chkconfig ] && /sbin/chkconfig --del ctdb
80 if [ "$1" -ge "1" ]; then
81 %{initdir}/ctdb restart >/dev/null 2>&1 || true
85 #######################################################################
87 #######################################################################
92 %config(noreplace) %{_sysconfdir}/sysconfig/ctdb
93 %config(noreplace) %{_sysconfdir}/ctdb/functions
94 %attr(755,root,root) %{initdir}/ctdb
96 %{_docdir}/ctdb/README.eventscripts
97 %{_docdir}/ctdb/recovery-process.txt
98 %{_sysconfdir}/ctdb/events.d/00.ctdb
99 %{_sysconfdir}/ctdb/events.d/10.interface
100 %{_sysconfdir}/ctdb/events.d/20.multipathd
101 %{_sysconfdir}/ctdb/events.d/40.vsftpd
102 %{_sysconfdir}/ctdb/events.d/41.httpd
103 %{_sysconfdir}/ctdb/events.d/50.samba
104 %{_sysconfdir}/ctdb/events.d/60.nfs
105 %{_sysconfdir}/ctdb/events.d/61.nfstickle
106 %{_sysconfdir}/ctdb/events.d/70.iscsi
107 %{_sysconfdir}/ctdb/events.d/90.ipmux
108 %{_sysconfdir}/ctdb/events.d/91.lvs
109 %{_sysconfdir}/ctdb/events.d/99.routing
110 %{_sysconfdir}/ctdb/statd-callout
114 %{_bindir}/ctdb_ipmux
115 %{_bindir}/ctdb_diagnostics
117 %{_mandir}/man1/ctdb.1.gz
118 %{_mandir}/man1/ctdbd.1.gz
119 %{_mandir}/man1/onnode.1.gz
120 %{_includedir}/ctdb.h
121 %{_includedir}/ctdb_private.h
124 * Thu Dec 3 2008 : Version 1.0.67
125 - Add a document describing the recovery process.
126 - Fix a bug in "ctdb setdebug" where it would refuse to set a negative
128 - Print the list of literals for debug names if an invalid one was given
130 - Redesign how "ctdb reloadnodes" works and reduce the amont of tcp teardowns
131 used during this event.
132 - Make it possible to delete a public ip from all nodes at once using
134 * Mon Nov 24 2008 : Version 1.0.66
135 - Allow to change the recmaster even when we are not frozen.
136 - Remove two redundant SAMBA_CHECK variables from the sysconf example
137 - After a node failure it can take very long before some lock operations
138 ctdb needs to perform are allowed/works with gpfs again. Workaround this
139 by treating a hang/timeout as success.
140 - Dont override CTDB_BASE is fet in the shell already
141 - Always send keepalive packets regardless of whether the link is idle or not.
142 - Rewrite the disable/enable flag update logic to prevent a race between
143 "ctdb disable/enable" and the recovery daemon when updating the flags to
145 * Thu Nov 13 2008 : Version 1.0.65
146 - Update the sysconfig example: The default debug level is 2 (NOTICE) and not
148 - Add support for a CTDB_SOCKET environment variable for the ctdb command
149 line tool. If set, this overrides the default socket the ctdb tool will
151 - Add logging of high latency operations.
152 * Mon Oct 22 2008 : Version 1.0.64
153 - Add a context and a timed event so that once we have been in recovery for
154 too long we drop all public addresses.
155 * Mon Oct 20 2008 : Version 1.0.63
156 - Remove logging of "periodic cleanup ..." in 50.samba
157 - When we reload a nodes file, we must detect this and reload the file also
158 in the recovery daemon before we try to dereference somethoung beyond the end
160 * Thu Oct 16 2008 : Version 1.0.62
161 - Allow multiple eventscritps using the same prefix number.
162 It is undefined which order scripts with the same prefix will execute in.
163 * Wed Oct 15 2008 : Version 1.0.61
164 - Use "route add -net" instead of "ip route add" when adding routes in 99.routing
165 - lower the loglevel os several debug statements
166 - check the status returned from ctdb_ctrl_get_tickles() before we try to print them out to the screen.
167 - install a new eventscript 20.multipathd whoich can be used to monitor that multipath devices are healthy
168 * Wed Oct 15 2008 : Version 1.0.60
169 - Verify that nodes we try to ban/unban are reachable and print an error othervise.
170 - 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.
171 - From Mathieu Parent : log debian systems log the package versions in ctdb diagnostics
172 - From Mathieu Parent : specify logdir location for debian (this patch was later reversed)
173 - From Michael Adams : allow # comments in nodes/public_addresses files
174 * Tue Oct 7 2008 : Version 1.0.59
175 - Updated "reloadnodes" logic. Instead of bouncing the entire tcp layer it is sufficient to just close and reopen all outgoing tcp connections.
176 - 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)
177 - IDR tree fix from Jim Houston
178 - Better handling of critical events if the local clock is suddenly changed forward by a lot.
179 - Fix three slow memory leaks in the recovery daemon
180 - New ctdb command : ctdb recmaster which prints the pnn of the recmaster
181 - Onnode enhancements from Martin S : "healthy" and "connected" are now possible nodespecifiers
182 - From Martin S : doc fixes
183 - lowering some debug levels for some nonvital informational messages
184 - Make the daemon daemon monitoring stronger and allow ctdbd to detect a hung
186 - From C Cowan : patches to compile ipv6 under AIX
187 - zero out some structs to keep valgrind happy
188 * Wed Aug 27 2008 : Version 1.0.58
189 - revert the name change tcp_tcp_client back to tcp_control_tcp so
191 - Updates to the init script from Abhijith Das <adas@redhat.com>
192 * Mon Aug 25 2008 : Version 1.0.57
193 - initial support for IPv6
194 * Mon Aug 11 2008 : Version 1.0.56
195 - fix a memory leak in the recovery daemon.
196 * Mon Aug 11 2008 : Version 1.0.55
197 - Fix the releaseip message we seond to samba.
198 * Fri Aug 8 2008 : Version 1.0.54
199 - fix a looping error in the transaction code
200 - provide a more detailed error code for persistent store errors
201 so clients can make more intelligent choices on how to try to recover
202 * Thu Aug 7 2008 : Version 1.0.53
203 - Remove the reclock.pnn file it can cause gpfs to fail to umount
204 - New transaction code
205 * Mon Aug 4 2008 : Version 1.0.52
206 - Send an explicit gratious arp when starting sending the tcp tickles.
207 - When doing failover, issue a killtcp to non-NFS/non-CIFS clients
208 so that they fail quickly. NFS and CIFS already fail and recover
210 - Update the test scripts to handle CTRL-C to kill off the test.
211 * Mon Jul 28 2008 : Version 1.0.51
212 - Strip off the vlan tag from bond devices before we check in /proc
213 if the interface is up or not.
214 - Use testparm in the background in the scripts to allow probing
215 that the shares do exist.
216 - Fix a bug in the logging code to handle multiline entries better
217 - Rename private elements from private to private_data
218 * Fri Jul 18 2008 : Version 1.0.50
219 - Dont assume that just because we can establish a TCP connection
220 that we are actually talking to a functioning ctdb daemon.
221 So dont mark the node as CONNECTED just because the tcp handshake
223 - Dont try to set the recmaster to ourself during elections for those
224 cases we know this will fail. To remove some annoying benign but scary
225 looking entries from the log.
226 - Bugfix for eventsystem for signal handling that could cause a node to
228 * Thu Jul 17 2008 : Version 1.0.49
229 - Update the safe persistent update fix to work with unpatched samba
231 * Thu Jul 17 2008 : Version 1.0.48
232 - Update the spec file.
233 - Do not start new user-triggered eventscripts if we are already
234 inside recovery mode.
235 - Add two new controls to start/cancel a persistent update.
236 A client such as samba can use these to tell ctdbd that it will soon
237 be writing directly to the persistent database tdb file. So if
238 samba is -9ed before it has eitehr done the persistent_store or
239 canceled the operation, ctdb knows that the persistent databases
240 'may' be out of sync and therefore a full blown recovery is called for.
241 - Add two new options :
242 CTDB_SAMBA_SKIP_CONF_CHECK and CTDB_SAMBA_CHECK_PORTS that can be used
243 to override what checks to do when monitoring samba health.
244 We can no longer use the smbstatus, net or testparm commands to check
245 if samba or its config is healthy since these commands may block
246 indefinitely and thus can not be used in scripts.
247 * Fri Jul 11 2008 : Version 1.0.47
248 - Fix a double free bug where if a user striggered (ctdb eventscript)
249 hung and while the timeout handler was being processed a new user
250 triggered eventscript was started we would free state twice.
251 - Rewrite of onnode and associated documentation.
252 * Thu Jul 10 2008 : Version 1.0.46
253 - Document both the LVS:cingle-ip-address and the REMOTE-NODE:wan-accelerator
255 - Add commands "ctdb pnn", "ctdb lvs", "ctdb lvsmaster".
256 - LVS improvements. LVS is the single-ip-address mode for a ctdb cluster.
257 - Fixes to supress rpmlint warnings
259 - Change \s to [[:space:]] in some scripts. Not all RHEL5 packages come
260 with a egrep that handles \s even same version but different arch.
261 - Revert the change to NFS restart. CTDB should NOT attempt to restart
263 - Rewrite of the waitpid() patch to use the eventsystem for handling
265 * Tue Jul 8 2008 : Version 1.0.45
266 - Try to restart the nfs service if it has failed to respond 3 times in a row.
267 - waitpid() can block if the child does not respond promptly to SIGTERM.
268 ignore all SIGCHILD signals by setting SIGCHLD to SIG_DEF.
269 get rid of all calls to waitpid().
270 - make handling of eventscripts hanging more liberal.
271 only consider the script to have failed and making the node unhealthy
272 IF the eventscript terminated wiht an error
273 OR the eventscript hung 5 or more times in a row
274 * Mon Jul 7 2008 : Version 1.0.44
275 - Add a CTDB_VALGRIND option to /etc/sysconfig/ctdb to make it start
276 ctdb under valgrind. Logs go to /var/log/ctdb_valgrind.PID
277 - Add a hack to show the control opcode that caused uninitialized data
278 in the valgrind output by encoding the opcode as the line number.
279 - Initialize structures and allocated memory in various places in
280 ctdb to make it valgrind-clean and remove all valgrind errors/warnings.
281 - If/when we destroy a lockwait child, also make sure we cancel any pending transactions
282 - If a transaction_commit fails, delete/cancel any pending transactions and
283 return an error instead of calling ctdb_fatal()
284 - When running ctdb under valgrind, make sure we run it with --nosetsched and also
285 ensure that we do not use mem-mapped i/o when accessing the tdb's.
286 - zero out ctdb->freeze_handle when we free/destroy a freeze-child.
287 This prevent a heap corruption/ctdb crash bug that could trigger
288 if the freeze child times out.
289 - we dont need to explicitely thaw the databases from the recovery daemon
290 since this is done implicitely when we restore the recovery mode back to normal.
291 - track when we start and stop a recovery. Add the 'time it took to complete the
292 recovery' to the 'ctdb uptime' output.
293 Ensure by tracking the start/stop recovery timestamps that we do not
294 check that the ip allocation is consistend from inside the recovery daemon
295 while a different node (recovery master) is performing a recovery.
296 This prevent a race that could cause a full recovery to trigger if the
297 'ctdb disable/enable' commands took very long.
298 - The freeze child indicates to the master daemon that all databases are locked
299 by writing data to the pipe shared with the master daemon.
300 This write sometimes fail and thus the master daemon never notices that the databases
301 are locked cvausing long timeouts and extra recoveries.
302 Check that the write is successful and try the write again if it failed.
303 - In each node, verify that the recmaster have the right node flags for us
304 and force a push of our flags to the recmaster if wrong.
305 * Tue Jul 1 2008 : Version 1.0.43
306 - Updates and bugfixes to the specfile to keep rpmlint happy
307 - Force a global flags update after each recovery event.
308 - Verify that the recmaster agrees with our node flags and update the
310 - When writing back to the parent from a freeze-child across the pipe,
311 loop over the write in case the write failed with an error othervise
312 the parent will never be notified tha the child has completed the operation.
313 - Automatically thaw all databases when recmaster marks us as being in normal
314 mode instead of recovery mode.
315 * Fri Jun 13 2008 : Version 1.0.42
316 - When event scripts have hung/timedout more than EventScriptBanCount times
317 in a row the node will ban itself.
318 - Many updates to persistent write tests and the test scripts.
319 * Wed May 28 2008 : Version 1.0.41
320 - Reactivate the safe writes to persistent databases and solve the
321 locking issues. Locking issues are solved the only possible way,
322 by using a child process to do the writes. Expensive and slow but... .
323 * Tue May 27 2008 : Version 1.0.40
324 - Read the samba sysconfig file from the 50.samba eventscript
325 - Fix some emmory hierarchical bugs in the persistent write handling
326 * Thu May 22 2008 : Version 1.0.39
327 - Moved a CTDB_MANAGES_NFS, CTDB_MANAGES_ISCSI and CTDB_MANAGES_CSFTPD
328 into /etc/sysconfig/ctdb
329 - Lowered some debug messages to not fill the logfile with entries
330 that normally occur in the default configuration.
331 * Fri May 16 2008 : Version 1.0.38
332 - Add machine readable output support to "ctdb getmonmode"
333 - Lots of tweaks and enhancements if the event scripts are "slow"
334 - Merge from tridge: an attempt to break the chicken-and-egg deadlock that
335 net conf introduces if used from an eventscript.
336 - Enhance tickles so we can tickle an ipv6 connection.
337 - Start adding ipv6 support : create a new container to replace sockaddr_in.
338 - Add a checksum routine for ipv6/tcp
339 - When starting up ctdb, let the init script do a tdbdump on all
340 persistent databases and verify that they are good (i.e. not corrupted).
341 - Try to use "safe transactions" when writing to a persistent database
342 that was opened with the TDB_NOSYNC flag. If we can get the transaction
343 thats great, if we cant we have to write anyway since we cant block here.
344 * Mon May 12 2008 : Version 1.0.37
345 - When we shutdown ctdb we close the transport down before we run the
346 "shutdown" eventscripts. If ctdb decides to send a packet to a remote node
347 after we have shutdown the transport but before we have shutdown ctdbd
348 itself this could lead to a SEGV instead of a clean shutdown. Fix.
349 - When using the "exportfs" command to extract which NFS export directories
350 to monitor, exportfs violates the "principle of least surprise" and
351 sometimes report a single export line as two lines of text output
352 causing the monitoring to fail.
353 * Fri May 9 2008 : Version 1.0.36
354 - fix a memory corruption bug that could cause the recovery daemon to crash.
355 - fix a bug with distributing public ip addresses during recovery.
356 If the node that is the recovery master did NOT use public addresses,
357 then it assumed that no other node in the cluster used them either and
358 thus skipped the entire step of reallocating public addresses.
359 * Wed May 7 2008 : Version 1.0.35
360 - During recovery, when we define the new set of lmasters (vnnmap)
361 only consider those nodes that have the can-be-lmaster capability
362 when we create the vnnmap. unless there are no nodes available which
363 supports this capability in which case we allow the recmaster to
364 become lmaster capable (temporarily).
365 - Extend the async framework so that we can use paralell async calls
366 to controls that return data.
367 - If we do not have the "can be recmaster" capability, make sure we will
368 lose any recmaster elections, unless there are no nodes available that
369 have the capability, in which case we "take/win" the election anyway.
370 - Close and reopen the reclock pnn file at regular intervals.
371 Make it a non-fatal event if we occasionally fail to open/read/write
373 - Monitor that the recovery daemon is still running from the main ctdb
374 daemon and shutdown the main daemon when recovery daemon has terminated.
375 - Add a "ctdb getcapabilities" command to read the capabilities off a node.
376 - Define two new capabilities : can be recmaster and can be lmaster
377 and default both capabilities to YES.
378 - Log denied tcp connection attempts with DEBUG_ERR and not DEBUG_WARNING
379 * Thu Apr 24 2008 : Version 1.0.34
380 - When deleting a public ip from a node, try to migrate the ip to a different
382 - Change catdb to produce output similar to tdbdump
383 - When adding a new public ip address, if this ip does not exist yet in
384 the cluster, then grab the ip on the local node and activate it.
385 - When a node disagrees with the recmaster on WHO is the recmaster, then
386 mark that node as a recovery culprit so it will eventually become
388 - Make ctdb eventscript support the -n all argument.
389 * Thu Apr 10 2008 : Version 1.0.33
390 - Add facilities to include site local adaptations to the eventscript
391 by /etc/ctdb/rc.local which will be read by all eventscripts.
392 - Add a "ctdb version" command.
393 - Secure the domain socket with proper permissions from Chris Cowan
394 - Bugfixes for AIX from Chris Cowan
395 * Wed Apr 02 2008 : Version 1.0.32
396 - Add a control to have a node execute the eventscripts with arbitrary
397 command line arguments.
398 - Add a control "rddumpmemory" that will dump the talloc memory allocations
399 for the recovery daemon.
400 - Decorate the talloc memdump to produce better and easier memory leak
402 - Update the RHEL5 iscsi tgtd scripts to allow one iscsi target for each
404 - Add two new controls "addip/delip" that can be used to add/remove public
405 addresses to a node at runtime. After using these controls a "ctdb recover"
406 ir required to make the changes take.
407 - Fix a couple of slow memory leaks.
408 * Tue Mar 25 2008 : Version 1.0.31
409 - Add back controls to disable/enable monitoring on a node.
410 - Fix a memory leak where we used to attach CALL data to the ctdb structure
411 when performing a local call. Memory which would be lost if the call was
413 - Reduce the loglevel for the log output when someone connects to a non
414 public ip address for samba.
415 - Redo and optimize the vacuuming process to send only one control to each
416 other node containing all records to be vacuumed instead of one
417 control per node per record.
418 * Tue Mar 04 2008 : Version 1.0.30
419 - Update documentation cor new commands and tuneables
420 - Add machinereadable output to the ip,uptime and getdebug commands
421 - Add a moveip command to manually failover/failback public ips
422 - Add NoIPFallback tuneable that prevents ip address failback
423 - Use file locking inside the CFS as alternative to verify when other nodes
424 Are connected/disconnected to be able to recover from split network
425 - Add DisableWhenUnhealthy tunable
426 - Add CTDB_START_AS_DISABLED sysconfig param
427 - Add --start-as-disabled flag to ctdb
428 - Add ability to monitor for OOM condition
429 * Thu Feb 21 2008 : Version 1.0.29
430 - Add a new command to make expansion of an existing cluster easier
431 - Fix bug with references to freed objects in the ctdb structure
432 - Propagate debuglevel changes to the recovery daemon
433 - Merge patches to event scripts from Mathieu Parent :
434 - MP: Simulate "service" on systems which do not provide this tool
435 - MP: Set correct permissions for events.d/README
436 - Add nice helper functions to start/stop nfs from the event scripts
437 * Fri Feb 08 2008 : Version 1.0.28
438 - Fix a problem where we tried to use ethtool on non-ethernet interfaces
439 - Warn if the ipvsadm packege is missing when LVS is used
440 - Dont use absolute pathnames in some of the event scripts
441 - Fix for persistent tdbs growing inifinitely.
442 * Wed Feb 06 2008 : Version 1.0.27
443 - Add eventscript for iscsi
444 * Thu Jan 31 2008 : Version 1.0.26
445 - Fix crashbug in tdb transaction code
446 * Tue Jan 29 2008 : Version 1.0.25
447 - added async recovery code
448 - make event scripts more portable
449 - fixed ctdb dumpmemory
450 - more efficient tdb allocation code
451 - improved machine readable ctdb status output
453 * Wed Jan 16 2008 : Version 1.0.24
454 - added syslog support
455 - documentation updates
456 * Wed Jan 16 2008 : Version 1.0.23
457 - fixed a memory leak in the recoveryd
458 - fixed a corruption bug in the new transaction code
459 - fixed a case where an packet for a disconnected client could be processed
460 - added http event script
461 - updated documentation
462 * Thu Jan 10 2008 : Version 1.0.22
463 - auto-run vacuum and repack ops
464 * Wed Jan 09 2008 : Version 1.0.21
465 - added ctdb vacuum and ctdb repack code
466 * Sun Jan 06 2008 : Version 1.0.20
467 - new transaction based recovery code
468 * Sat Jan 05 2008 : Version 1.0.19
469 - fixed non-master bug
470 - big speedup in recovery for large databases
471 - lots of changes to improve tdb and ctdb for high churn databases
472 * Thu Dec 27 2007 : Version 1.0.18
473 - fixed crash bug in monitor_handler
474 * Tue Dec 04 2007 : Version 1.0.17
475 - fixed bugs related to ban/unban of nodes
476 - fixed a race condition that could lead to monitoring being permanently disabled,
477 which would lead to long recovery times
478 - make deterministic IPs the default
479 - fixed a bug related to continuous recovery
480 - added a debugging option --node-ip