PR pkg/48465 by ISIHARA Takanori.
Version 5.6
IMPROVEMENTS:
* SMTP AUTH LOGIN support added (MS Exchange SMTP authentication
should now work).
* favicon.ico added to the HTTP interface.
BUGFIXES:
* If an undefined checksum test was used and the file did not exist
on Monit start, Monit would return an error.
* If the configuration file ended with a comment but with no trailing LF
character, Monit would return syntax error.
* If a service timed out after too many restarts and alert was used as
the action, then the Timeout flag remained set even if the service
recovered.
* SmartOS zone system memory usage report fix.
* Escape mail messages properly for sending via SMTP.
* Escape XML messages properly.
* Compilation: fix the configure script to support default compiler
paths when searching for OpenSSL (fixes library search on multi-
architecture platforms like Debian and Ubuntu).
Version 5.5.1
IMPROVEMENTS:
* Info and debug messages are no longer sent to stderr, only to stdout.
Thanks to Sergey Kirpitchev for initial patch.
* Improved output from 'check program', If the program returns an error
message, include only that message in alert $DESCRIPTION so users can
compose their own alert format. If program provided no output on
error, use a default message.
* Improved "check system", $HOST can now be used as a service name.
$HOST will expand to the system hostname. Example: check system $HOST
BUGFIXES:
* Fixed "Unable to read magic" which was reported on first Monit start.
Version 5.5
IMPROVEMENTS:
* check program:
- Multiple exit values can be tested within single program check
- Exit value test supports multiple cycles option ("for X cycles")
- If exit value test matches and the stderr has no data, try stdout
Example syntax:
check program mytest with path "/usr/bin/mytest.sh" with timeout 1000 seconds
if status == 2 then exec "/usr/local/bin/fix_script.sh"
if status == 2 for 6 cycles then unmonitor
if status == 10 then alert
* Renamed mail header (message-id and mime-version) to prevent
triggering spam check of capitalization. Thanks to Ryan Lee
for tips.
* The 'check system <name>' statement sets the system hostname in mail
alerts and initial hostname in M/Monit.
* Increase the default mailserver timeout to 30 seconds.
* Add support for OpenBSD 5.x
BUGFIXES:
* Fix the rare hung on linux which may occur during program execution.
Thanks to Nick Upson for report.
* In the case that the process start/restart execution failed,
monit kept "Execution failed" flag even if the process was
recovered later (for example it was starting slowly or manually
recovered).
* Fix the mail alert (strict SMTP implementation) to pass
MTA-side sanity checks like postscreen. Thanks to Len Conrad
for report.
* The -t option tests the configuration file syntax even if the
file permissions are wrong. Thanks to Adam Nielsen for report.
* Do not display the default non-existence test for the check
program (not applicable in the check program context).
Version 5.4
IMPROVEMENTS:
* New process uptime test added. Allows to do some action in
the case that the process uptime matches the given limit.
For example to restart the process once per 3 days:
if uptime > 3 days then restart
* Linux uCLibc support: use internal getloadavg implementation
in the case that the system libc doesn't implement it.
BUGFIXES:
* The monit hostname will fallback to plain machine's hostname if the
lookup for FQDN hostname didn't found matching entry. The problem
was, that based on the order in the /etc/hosts the FQDN lookup
returned sometimes 'localhost' instead of the FQDN hostname.
* The CPU usage for multi-threaded processes on multi-core machine was
reported incorrectly in the case that the process used more CPU
resources then equivalent to one core. Thanks to Tom Pepper for patch.
* The content match test now sends one event per cycle and pattern.
Even if there are multiple lines matching the same pattern, only
one event will be generated. Also the event rate is fixed now, so
it is possible to require match for X cycles before generating the
event.
* The /proc/ files content match test was skipped, as the file size on
the procfs is 0, so monit supposed that there is no content to read.
* FreeBSD: If the monitored process had children with multiple threads,
the total memory usage was reported incorrectly. Thanks to Phil Kulin
for reporting the problem.
* Allow reading status and perform Monit actions when using client SSL
certificate. Previously, if Monit http server was setup to use ssl
and a client cert, status and action failed. Thanks to Markus Linnala
for initial patch.
* When the process is starting/stopping, do the process state check more
effectively to not stress the low power devices with aggresive polling.
Thanks to Thomas Petazzoni for initial patch.
* Make the process start/stop wait resistant to large time changes.
* Compilation: If PAM is enabled but the PAM headers or library are not
found by the configure script, it will report error.
* Cross-compilation: the configure checks the setjmp and vsnprintf with
test program which usually cannot be executed when cross-compiling
for other architecture. The configure script now takes the following
arguments which allow to specify whether the setjmp works on this
platform and whether the vsnprintf is C99 compliant. Thanks to
Thomas Petazzoni for patch.
./configure \
libmonit_cv_setjmp_available=[yes|no] \
libmonit_cv_vsnprintf_c99_conformant=[yes|no]
* Manual page language fixes. Thanks to Jonathan Boulle for patch.
Version 5.3.2
BUGFIXES:
* Fix bug #34801: The file content match test did reset of the
read position in the case that the unmonitor or stop action
was done. When the file monitoring was enabled again, the
content match test was applied to the content which was
tested already.
* Log error details in the case that the name resolving failed.
* Fix the system cpu usage statistics when pattern based process
check is used and the service is restarted. Thanks to Wayne
Lawrence for report.
* AIX 6.1 compile fix. Thanks to Benedikt Wegmann for patch.
* Debian Bug#652715: "include files not found" warning. Do not
display the warning if the include directory is empty.
Version 5.3.1
IMPROVEMENTS:
* Log the particular connection attempt failure in debug mode
when the retry is enabled.
* Monit can deliver events and status to independent M/Monit
instances if multiple mmonit URLs are set:
set mmonit https://user1:pass1@mmonit1/collectorhttps://user2:pass2@mmonit2/collector
BUGFIXES:
* The ICMP echo (ping) test may report false positive error
if the machine where Monit is running has heavy ICMP
traffic generated by other applications.
* The file content match test will be performed even on the
existing content when Monit starts. The last position is
saved to the statefile, so monit won't generate alert
after restart. Note that when you start the monit 5.3.1
the first time, it can do actions for content match which
was handled by previous monit version already as the
previous monit versions didn't saved the position.
* Make the monitoring state persistent for manual mode services.
* Display the memory usage total % in the status overview.
The memory usage in kB displayed the total already, so the
percentage didn't match.
* Fix the HTML overview page alignment in the Internet Explorer.
Thanks to Darhl Thomason for patch.
* Extend the SSL library search path for Debian Sid.
* Fix Solaris 10 compilation and Sun Studio support.
* Fix sporadic SSL routines:func(169):reason(161) errors
* If MySQL protocol test failed, report the correct MySQL
error code. Thanks to Vitaly Lipatov for patch.
Version 5.3
* New 'check program' statement added. Allows to check the exit
status of an external program or script from Monit.
* Added crontab style support for individual services. You can
now specify when an individual service should run its checks
(or not run). You can now, for instance, specify that apache
should be checked continuously, except between 1AM-5AM on
Sunday.
* Connection retry option added. Allows to retry a network
connection in the same testing cycle before reporting an error.
* Detailed protocol connection errors are now included in alerts.
* The HTML overview page displays the CPU and memory total now
(including children), so real service related usage is displayed
also for services which spawn worker processes, such as Apache
or Spamassassin.
* HTML view improvements
* Fix MySQL protocol test: MySQL 5.5.12 returns new error code in
the case of authentication failure.
* Fix Debian bug #621047: monit fails to build after SSLv2 removal
* Fix crash on Solaris which may occur if the system load is zero.
Thanks to Paul Sun for report.
* The stacktrace logging on error is disabled in -v (verbose) mode
as it was too verbose for common service debugging tasks, it can
be enabled using -vv option.
* Improve how fast Monit check if a program was started or stopped.
Thanks to Michael Renner for patch.
* Fix the monitoring state presentation during service restart which
temporarily displayed "Not monitored", whereas the monitoring was
enabled.
* The "data collected" is updated only if the check was not skipped.
Version 5.2.5
* Fix process match check - when the monitored process failed and
was restarted by Monit, Monit didn't recognized it is running
after the restart and reported start failure (similar on stop).
Thanks to Kenichi Futatsumori for report and helping to root
cause the problem.
* Fix Debian #617259: symbolic links in the filesystem check doesn't
work. Thanks to Sergey B Kirpichev for report.
* Fix Debian bug #614984: smtp protocol test issues both EHLO and
HELO. Thanks to Sergey B Kirpichev for report.
* Fix bug #32583: Multiple SIP OPTIONS messages use the same header
data. Thanks to Hugh Waite for patch.
* Try harder to get FQDN hostname for the host where monit is running.
The hostname in the $HOST variable which is used in the mail sender
may thus change. Thanks to Sergey B Kirpichev for patch.
* AIX: Fix the time display which was off by GMT difference. Thanks
to Helen Chen for report.
* AIX: Fix the M/Monit heartbeat. Thanks to Helen Chen for report.
* Support symbolic link to monit configuration file.
* Fix crash when monit daemon start delay option was used and monit
was signalized to stop before the start delay passed. Thanks to
John Schult for report.
Version 5.2.4
NEW FEATURES AND FUNCTIONS:
* Added the "procmatch" CLI command which allows for easy testing
of pattern for process match check. The command takes regular
expression as an argument and displays all running processes
matching the pattern. Example usage:
$ monit procmatch "iChatAgent"
* Set the default log file mask to 0640 (originally it was 0664).
Thanks to Sergey B Kirpichev.
* Reduced monit memory footprint by ca. 10%.
BUGFIXES:
* FreeBSD, NetBSD, OpenBSD, MacOSX, Solaris filesystem check fix:
If block/character device was used in the filesystem path instead
of mountpoint, monit reported usage of wrong filesystem.
* NetBSD filesystem check: Fix space usage report.
* Fix memory usage monitoring in OpenVZ VPS 2.6.32 virtual hosts.
Thanks to Kelly for report.
* If the protocol test failed, show the request in the event. Thanks
to Marco for report.
* Randomize the mail message id to prevent duplicates in the case, that
the same hostname is used on multiple hosts running monit and messages
are generated in the same second in parallel. Thanks to Sergey B
Kirpichev.
* Spelling fixes. Thanks to Sergey B Kirpichev.
Version 5.2.3
BUGFIXES:
* Mysql protocol test supports mysql 5.5.x and newer now.
Version 5.2.2
BUGFIXES:
* Fix crash on MacOSX
* ICMP echo test (ping):
- bug #31128: do not log error if different response type is received
- bug #31129: do not require root to use ping test. Privilege to create
raw socket is still required, but on some platforms such as Solaris it
can be granted to non-root users too. If the user has no permission to
perform ping, monit will skip the icmp test and log message (in debug
mode only).
* rsync protocol test:
- wait for full server response and verify exit was received
- bug #31249: send full version to rsync server. Thanks to John Hall
for report
Version 5.2.1
BUGFIXES:
* HTTP and URL protocol tests: Fixed a problem where HTTP protocol
tests using a specific request always failed. This bug may also
affect URL tests. The problem was caused by faulty URL encoding. In
the process of fixing this bug the new feature that allowed slash in
service names has been reverted and instead will be added in a later
release.
Version 5.2
NEW FEATURES AND FUNCTIONS:
* Added support for monitoring processes without pidfile using pattern
matching. You can use POSIX regular expressions or string matching
process name with arguments as provided by the 'ps' utility. If the
pattern matches multiple processes, the first match is used.
Example:
check process debian
matching "/usr/lib/vmware/bin/vmware-vmx .*deb.vmx"
* Added support for swap monitoring. Example:
check system myserver
if swap usage > 25% then alert
* Allow to override the default action when service doesn't exist. The
default action is restart, it can be customized with following
statement:
if [does] not exist [[<x> times within] <y> cycles] then <action1>
* Monit automatically registers credentials with M/Monit now, so it's
not necessary to set it manually in M/Monit anymore. To disable
credentials registration:
set mmonit https://monit:monit@10.0.0.1:8443/collector
and register without credentials
* Added memcache protocol test. Thanks to Sébastien Debrard for the
patch.
* Added openssl FIPS to Monit httpd. Thanks to Lior Okman for the
patch.
* The 'check system' can now use start/stop program statements too.
* Added the option to set the "Reply-To" mail header in mail-format.
* Display backtrace on error if debug mode is enabled (requires
backtrace support in libc)
BUGFIXES:
* Show real process uptime - formerly the presented uptime was based
on create and modify timestamp of process' pidfile which provides
invalid uptime if the pidfile is replaced and process keeps running
with original PID. Thanks to Nima Chavooshi for report.
* When user triggered action for some service (such as stop) and
before that action completed user triggered another action for the
same service (such as start), the second action has been ignored.
Monit will not accept new action and return temporary error until
the previous action completed.
* If process resource usage gathering failed, retry next cycle as the
error can be temporary.
* Fixed sporadic failures when SSL was used.
* ICMP echo test (ping):
- fixed sporadic false positive/negative
- removed limit of 20 pings per cycle
* DNS test:
- accept NS root request refusal as correct response because
server reacts on request
- accept authority answer as alternative to record. Thanks to
Nick Osborn for patch
* RADIUS test fix. Thanks to Alan DeKok for patch.
* M/Monit heartbeat is fully independent of testing cycle now to
prevent false positive when service test blocks.
* Fixed SMTP STARTTLS protocol, required for servers that adhere
strictly to RFC 3207 4.2. Thanks to Lorenzo A. Sedano Cadinanos for
patch.
* Service name:
- allow the service name to start with "/"
- fixed handling of the service names which contain "/" in the
name in Monit web interface. Thanks to Artyom Khafizov for
patch.
* When 'check system' is not defined, monit adds it automatically
using hostname for service name. If existing service was defined
with the same service name (matching hostname), monit didn't added
the entry and reported confusing error message pointing to the end
of configuration file. Thanks to Thorsten Kampe for report and help.
* Remove extra NL characters from message when resource succeeded
event is sent. The extra NL character may break the mail headers.
Thanks to Hanno Boeck for patch.
* Fixed display of cpu user/system/wait usage which temporarily
displayed -1.0% between two monitoring cycles while cpu monitoring
was initializing. Thanks to Marcus Muelbuesch for report.
* Fixed display of port response time as -1 if 'monit status' was
called in the middle of service test.
* Fixed display of service initializing state after monit start or
reload.
* Fixed MONIT_DESCRIPTION environment variable. Thanks to Marco
Roeland for patch
* AIX:
- fixed compilation
- fixed system load average monitoring
- fixed ICMP echo test
* Mac OS X:
- allow monitoring of system-wide load average, cpu and memory
usage even if
Monit is running as non-root user
* NetBSD:
- fixed ICMP echo test
Version 5.1.1
BUGFIXES:
* Fix FTP protocol test. Thanks to Axel Reinhold for report.
* Fix the HTTP protocol test's hostheader option which was added in 5.1.
Thanks to Naoya Nakazawa for report.
* Removed warning about missing system service check. Missing system service
check is not error and it shouldn't be reported as such.
* Fix manual page formating. Thanks to Stefan Alfredsson for report.
Version 5.1
NEW FEATURES AND FUNCTIONS:
* It is now possible to define any action for the restart timeout rule.
Multiple restart timeout rules can also be defined. Example:
if 3 restarts within 5 cycles then exec "/foo/bar"
if 8 restarts within 10 cycles then unmonitor
* Service can be added to multiple groups. Thanks to Brad Gessler
for suggestion. Syntax:
check filesystem wwwdata with path /www
group www
group filesystem
* Added GPS protocol test. Thanks to Sebastien Debrard for patch.
* Added RADIUS protocol test. Thanks to Alan DeKok for patch. Example syntax:
check process radiusd with pidfile /var/run/radiusd.pid
start program = "/etc/init.d/freeradius start"
stop program = "/etc/init.d/freeradius stop"
if failed
host 127.0.0.1 port 2000 type udp protocol radius secret testing1234
then alert
if 5 restarts within 5 cycles then timeout
* The HTTP protocol test now supports a hostheader option which allows to
override Host header in HTTP request. It can be used for example
to test a farm of HTTP servers by IP addresses and to set specific
Host header. Thanks to Brady Catherman for patch. Example:
if failed host 192.168.1.100 protocol http hostheader "example.com" then alert
* If an error occur during Monit command-line execution, report the error
and exit with 1, so it is possible to react if Monit is used from a script.
On success, 0 is returned as usual. Previously, Monit always exited with
0 even if an error occurred.
* Do not require SSL version type when specifying SSL communication with M/Monit
(SSL version is set to auto).
* If the Monit http interface failed to start, provide more details about
the reason.
BUGFIXES:
* Support resource monitoring (cpu usage, etc.) when Monit is running
inside virtual environment. Tested on:
- FreeBSD jail
- Solaris zone
- Linux Vserver
* Fix#26752: inside Solaris Zone, Monit failed to detect children
and computed host memory wrong
* On Solaris, FreeBSD, NetBSD and OpenBSD, Monit no longer needs to run as root user
in order to be able to watch process resource usage (cpu and memory).
* Send heartbeat to M/Monit even if Monit is busy in a long testing cycle to prevent
false alerts about non-responsive Monit agent.
* Fixed SMTP protocol test which may sometimes incorrectly
report ESMTP protocol failure. Thanks to Axel Reinhold for
report.
* Fixed content match check which reported only first
match during the same cycle. Thanks to Pavel Shevaev for
report.
* Allow for the use of complete SSL certificate chains.
Thanks to Lawrence Tan for patch.
* Added support for multiline greetings to FTP protocol test.
Thanks to Giovanni D'Cristina for report.
* Fix Debian Bug #541139: uses gethostbyname() and thus does
not work with "options inet6" in /etc/resolv.conf. Thanks to
Michael Stapelberg for patch.
* If Monit configuration allowed http interface access for a read-only
user and it was specified as the first allow entry, Monit command line
commands failed because it used the read-only account so commands
like start, stop, etc. were rejected. Monit will now use full access
regardless of allow option order. Thanks to Thorsten Kampe for report.
* Passive monitoring mode fixed. Thanks to Nelson Vale for report.
* Fixed#27784: wait_start/wait_stop can advance too quickly.
Thanks to Randy Puro for report.
* Solaris resource usage fixed when Monit was compiled with optimizations enabled.
* Fixed#28369: escape XML properly
* Check service name uniqueness when 'check system' is missing in monitrc and virtual
system service with name set to local hostname is added. Thanks to Marcus Muelbuesch
for report.
* Fix crash when queued event delivery was retried for service which was no longer
configured in Monit.
Version 5.0.3
BUGFIXES:
* Fixed#26664: crash on service timeout or unmonitor action
(introduced in 5.0.2). Thanks to Bretislav Kubesa and
Michael Shigorin for report.
* Removed the configure --without-resource option. If the user
who is running Monit doesn't have permissions to check the
processes state, the related checks are disabled dynamically.
Version 5.0.2
BUGFIXES:
* 35 improvements based on code scan with Klocwork
(http://www.klocwork.com/) which we were evaluating.
Huge thanks to Klocwork for their great product.
* Fixed#26382: if start or stop script for some service didn't
exist, monit logged error during configuration file parsing and
refused to start. Monit now just logs warning and continues.
Version 5.0.1
BUGFIXES:
* Fixed a bug where Monit did not stop logging succeeded events.
This bug occurred if PID, PPID, timestamp or size change tests
were used and failed and then succeeded again.
Version 5.0
NEW FEATURES AND FUNCTIONS:
* M/Monit support added. If you run Monit on more than one
server, you can use M/Monit to manage and control all your
Monit enabled servers from one simple Web Interface. See
http://mmonit.com/ for details.
* Support use of symbolic links in filesystem check. Thanks to
Aleksander Kamenik for suggestion. Example:
check filesystem rootfs path
/dev/disk/by-uuid/4ef973f7-67d1-4bb0-8223-cb1c692b72e4
if space usage > 95% then alert
if inode usage > 95% then alert
* If no 'set mailserver' was defined in monitrc, Monit tried to
fallback to localhost:25 SMTP server. This fallback was removed
since it may be confusing. If you want to deliver mail alerts
from Monit, the 'set mailserver' option is necessary. In case
it is missing, Monit will log appropriate error and hint to add
it.
* The generic send/expect protocol test limited the expect input
to 256 bytes. It's possible to set the input buffer for expect
globally - for example: set expectbuffer 20 kb Thanks to Asil
Carlin for suggestion.
* The following event types were added CONTENT, FSFLAGS, PID and
PPID and the following generic event types CHANGED and MATCH were
removed and replaced by the above types and with the existing SIZE,
CHECKSUM, TIMESTAMP events so the information is more specific
The event types are internal to Monit and unless you have used
either CHANGED or MATCH event in your alert filters, no change
is necessary (alerts are delivered as usual, the tests just use
different types internally).
* Monit now generates a unique id on first start and store the id
in a permanent file. This id is used in protocol communication
between Monit and M/Monit to pair a Monit instance with it's
host entry in M/Monit. By default the id file is placed in
$HOME/.monit.id. The location can be changed by using the set
idfile statement, for example:
set idfile /var/monit.id
* Monit now keep its service monitoring state even on Monit
restart. Previously Monit dropped the state when it was stopped
correctly. Services in manual monitoring mode will remember the
monitoring state across Monit restarts. If Monit is used in a
cluster, it is recommended to place the state file in a
temporary filesystem incase the primary machine will crash and
the the spare machine takeover, the state will be dropped on
reboot for the crashed machine and the services in manual
monitoring mode won't be started on reboot. For example the
"set statefile /tmp/monit.state" can be used to place the state
file in the /tmp/ filesystem.
* Added a protocol test for testing the LMTP protocol. Thanks
to Fco. Javier Felix for patch.
* Added the start delay option for daemon statement which allows
to pause Monit on its startup for a while. If monitored
services are started by init scripts in parallel on system
boot, Monit may be too fast and detect that the service is not
running (yet) and restart the service. Note that it's still
recommended Monit is setup to be responsible for service
startup (that is, don't use init to start Monit controlled
services, instead use Monit). This will ensure correct startup
without need for a start delay since Monit will have full
control of service startup. Many users start services from init
on boot anyway, so in such cases this option will solve their
problems. Default start delay is 0 which corresponds to the
current behavior. Example syntax which will make Monit wait one
minute before starting its first monitoring cycle:
--8<--
set daemon 5 with start delay 60
--8<--
Thanks to Fco. Javier Felix for patch.
* Added PAM support for Monit http interface authentication. Note
that PAM is not supported on all platforms - currently works on
Linux, Mac OS X, FreeBSD, NetBSD. Monit uses the PAM service "monit".
Here is a Monit PAM service example for Mac OS X which is able
to authenticate system users for Monit access -
/etc/pam.d/monit:
--8<--
# monit: auth account password session
auth sufficient pam_securityserver.so
auth sufficient pam_unix.so
auth required pam_deny.so
account required pam_permit.so
--8<--
And configuration for monitrc which allows only group admins
to access the http interface:
--8<--
set httpd port 2812 allow @admin
--8<--
See the PAM manual page for details on how to configure the PAM
service on your system and the available PAM plugins. Thanks to
Wilhelm Meier for patch.
* Added more detailed reports for Monit resource tests on service
recovery. Thanks to Lars Kotthoff for patch.
* Set locale to C.
* Added a protocol test for testing the SIP protocol which is
used by popular communication servers such as Asterisk and
FreeSWITCH. We received two patches for this protocol and have
taken code from both and merged them. Many thanks to Bret
McDanel and to Pierrick Grasland for supplying the patches.
* Added MONIT_DESCRIPTION to the list of environment variables
available to programs started by monit. Thanks to Morten
Bressendorff Schmidt for patch.
* If a service group is specified for Monit CLI action,
Monit no longer requires the "all" verb, so the following
command is possible:
monit -g web stop
If group is not specified (i.e. the -g option is omitted), the
service name or "all" is still required as a safeguard.
* Added an option to the 'set mailserver' statement so it is
possible to override the hostname used in SMTP EHLO/HELO and in
the Message-ID header when sending mail. Monit defaults to use
the localhost name. I.e. what you get when executing this
command 'uname -n'. Overriding the host name can be useful if
the host does not have a DNS entry and if the receiving
mailserver uses DNS verification as spam protection. The new
override option is:
set mailserver foo.bar.baz using hostname "my.monit.host"
* A new Event_Action type was added which reports actions
performed on Monit's administrator request (either via web
interface or CLI). If you don't want to received these events,
you can set the mail-filter for "action" event type.
* NOTA BENE: Monit start action is synchronous now. This improves
the startup sequence for dependent services, since Monit will
wait for parent service to start before trying to start the
child.
* It is now possible to define execution timeout for start and
stop commands. That is, how long Monit will wait after
executing a command before it assume execution failed. If the
timeout option is omitted, Monit defaults to 30 seconds. You
can override the timeout for example for services which are
starting slower.
Example syntax:
start program = "/bin/foo start" with timeout 60 seconds
* The event passed state is renamed to succeeded as this name
more reflects the state of things.
* The device service test is renamed to filesystem.
BUGFIXES:
* Some linux virtualization platforms report CPU count as 0.
Monit then dynamically disabled CPU usage monitoring. In such
case we now override the CPU count from 0 to 1 so resource
usage monitoring can continue. Thanks to Jenny Hopkins for
report.
* Increased the server socket backlog queue which will make Monit
able to handle more services. Thanks to Jochen Kramer.
* Fixed#24866: Email messages such as: cpu wait usage check
succeeded [current cpu wait usage=17.4%] were displayed as
"...usage<SOMEGARBAGE>.4%". The problem was incorrect transfer
encoding header in the email (the body itself was OK). Thanks
to Dave Cheney for report.
* When a Monit shutdown requested was issued while Monit were
working and testing services, Monit did not shutdown until all
work were done, i.e. until all services were tested. Monit will
now shutdown faster - as soon as it finish testing the current
service.
* Monit blocked/unblocked SIGTERM, SIGINT SIGHUP and SIGUSR1
signals during operation to protect certain code sections. When
a signal was sent during such a time, for example to stop
Monit, it was dropped and had to be retried in order to stop
Monit. This limitation is now removed and signals will be
processed at any time. Thanks to Nicola Tiling for report.
* If the Monit httpd allow option did not include a
user:password, Monit CLI logged the following error (even if
the action was performed anyway):
Cleartext credentials needed for basic authorization!
This error was false - even access restriction based on
host/net is sufficient - user and password is just one of
possible options (not requirement). Thanks to Gilad Benjamini
for report.
* Allow localhost as a value for the host header in the http
protocol test instead of setting an empty host header and let
the http server decide
* The 'if changed checksum ...' test can now be used even if a
monitored file doesn't exist at Monit startup. Thanks to Joe
Shang for report.
* If both event handlers (M/Monit and mail alerts) temporarily
failed at once and event queue was enabled, events will be
stored in the queue and delivered in the next cycle. However, a
bug caused delivery to be retried for every cycle for both
handlers if just one of them was recovered. Monit could then
deliver the same message multiple times until both handlers
recovered. The problem is now fixed and only one copy of the
event is sent even if only one handler did recover.
* Make unit in size test optional and default to byte unless
specified. So it is possible to write, if size > 1000 then ..
* Fixed handling of invalid input files in event queue handler.
Thanks to Fco.Javier Felix for patch.
* Set the content type to text/html for Monit web interface POST
responses. Thanks to Rich Drummond for patch.
* Fixed#23530: configure script will return error if bison,
byacc or yacc are not found at Monit compile time.
* Fix CPU and memory monitoring on Solaris (it was disabled on
Monit start)
* AIX fixes and extensions, Monit should run on AIX without
problems, including cpu, memory and filesystem monitoring
(tested with AIX 5.3). Thanks to Brian Downey for support
and help.
* HP-UX fixes and extensions, Monit should run on HP-UX without
problems, including cpu, memory and filesystem monitoring.
Thanks to Brian Downey for support and help.
* Fixed#23467: Don't exit, only issue a warning if the "include"
statement did not find any files to include.
* Fixed#23530: Event queue did not work with the default
unlimited slots.
* Fixed#23617: The process cpu usage is initializing in the
first cycle so the value is set to 0% - if the 'cpu usage <
xyz%' test was used to check that the process usage is higher
then given level, it was always true. Monit now skips the
process cpu usage check in the first cycle.
* Make sure Monit alerts has a unique message id. Thanks to Steve
Purcell for report
* Fixed possible crash when Monit is watching VPS environment on
Linux which reports number of CPUs as 0. Thanks to Marius
Schmidt for report.
* Cleanup event states during a service stop/unmonitor so old
events are not sent when the service is started/monitored again.
* Fixed#21989: Monit could start two instances of the process
when service restart is performed and the process is starting
slowly. Thanks to Nick Upson, Aaron Scamehorn and David Greaves
for report.
* Fixed#21550: Fix crash when Monit event queue contained an
empty file. Thanks to Douglas J Hunley for report.
* Fixed possible crash when the 'if changed checksum' test was
used along with restart action. Thanks to Brian Candler for
report.
* Fixed#22075: Allow using a mail address as username when using
SMTP authentication.
* Fixed#22191 and #19823: If the file content test does not match
anymore, reset the service error state. (Previous versions did
not clear the error state and kept showing a match in the status
listing and in the http interface).
* The 'if changed size ...' test can now be used even if the
monitored file does not exist on monit's start.
* If a htpasswd file is used to control Monit http interface
access and the hash type is set to MD5 but the file contains
wrong format (non-MD5), report the error and keep running.
Formerly Monit exited with an assert exception. Thanks to
Adrian Bridgett for report.
BACKWARD INCOMPATIBLE CHANGES:
* The current CPU usage test which checked the cpu usage of the
process itself plus the cpu usage of child processes was
renamed to TOTALCPU (otherwise it works the same). The new CPU
usage test checks the CPU usage of the process itself only.
This change was introduced to align the syntax with MEMORY and
TOTALMEMORY tests and to allow to test the CPU usage of
processes which fork child processes but the user don't want to
include children (such as Mythtv). Users who are using the CPU
check for services like Apache webserver to watch total cpu
utilization (including children) should rename the CPU
statement in their configuration to TOTALCPU.
There have been 195 commits by 32 people in the 10 months since 8.21
Executive summary: 8.22 is mainly a bug fix and performance improvement release.
tail(1) has fixes for handling initially non existent files, and df now has
better handling of specified disk device nodes. On the performance side,
there have been many improvements to the memory usage of various tools, which
are detailed at http://www.pixelbeat.org/programming/avoiding_large_buffers.html
There have been large throughput improvements in cut(1) and base64 encoding.
Also the md5sum and sha*sum utilities have support for using architecture
specific routines to improve performance, if allowed on your distribution.
There are also new features available, including SELinux "restorecon" support
in various file creation utilities through the -Z option, and shred has a new
--repeat option to continually select random items from the input.
See the NEWS below for more details.
Thanks to everyone who has contributed!
The following people contributed changes to this release:
Andreas Mohr (1)
Anton Ovchinnikov (2)
Assaf Gordon (8)
Benno Schulenberg (2)
Bernhard Voelker (28)
Cojocaru Alexandru (4)
Colin Leitner (1)
D. Hugh Redelmeier (1)
Daniel J Walsh (1)
Enrico Scholz (1)
Eric Blake (1)
FUJIWARA Katsunori (1)
Filipus Klutiero (1)
Gian Piero Carrubba (1)
Jarkko Sakkinen (4)
Javier López (1)
Jim Meyering (6)
Joachim Schmitz (2)
John (1)
Karl Berry (1)
Ken Booth (1)
Mike Frysinger (1)
Ondrej Oprala (2)
Ondřej Vašík (2)
Paul Eggert (10)
Pádraig Brady (106)
Rasmus Villemoes (1)
Rémy Lefevre (1)
Sergio Durigan Junior (1)
Stefano Lattarini (3)
Tiger Lee (1)
Torbjörn Granlund (1)
Special thanks to Bernhard Voelker for his many considered patches and reviews,
and to Assaf Gordon for his work on new (and future) features.
Pádraig [on behalf of the coreutils maintainers]
==================================================================
Here is the GNU coreutils home page:
http://gnu.org/s/coreutils/
For a summary of changes and contributors, see:
http://git.sv.gnu.org/gitweb/?p=coreutils.git;a=shortlog;h=v8.22
or run this command from a git-cloned coreutils directory:
git shortlog v8.21..v8.22
To summarize the 255 gnulib-related changes, run these commands
From a git-cloned coreutils directory:
git checkout v8.22
git submodule summary v8.21
==================================================================
Here are the compressed sources and a GPG detached signature[*]:
http://ftp.gnu.org/gnu/coreutils/coreutils-8.22.tar.xzhttp://ftp.gnu.org/gnu/coreutils/coreutils-8.22.tar.xz.sig
Use a mirror for higher download bandwidth:
http://ftpmirror.gnu.org/coreutils/coreutils-8.22.tar.xzhttp://ftpmirror.gnu.org/coreutils/coreutils-8.22.tar.xz.sig
[*] Use a .sig file to verify that the corresponding file (without the
.sig suffix) is intact. First, be sure to download both the .sig file
and the corresponding tarball. Then, run a command like this:
gpg --verify coreutils-8.22.tar.xz.sig
If that command fails because you don't have the required public key,
then run this command to import it:
gpg --keyserver keys.gnupg.net --recv-keys DF6FD971306037D9
and rerun the 'gpg --verify' command.
This release was bootstrapped with the following tools:
Autoconf 2.69
Automake 1.13.4
Gnulib v0.1-38-g0658e50
Bison 2.7
NEWS
* Noteworthy changes in release 8.22 (2013-12-13) [stable]
** Bug fixes
df now processes the mount list correctly in the presence of unstatable
mount points. Previously it may have failed to output some mount points.
[bug introduced in coreutils-8.21]
df now processes symbolic links and relative paths to special files containing
a mounted file system correctly. Previously df displayed the statistics about
the file system the file is stored on rather than the one inside.
[This bug was present in "the beginning".]
df now processes disk device nodes correctly in the presence of bind mounts.
Now df shows the base mounted file system rather than the last one mounted.
[This bug was present in "the beginning".]
install now removes the target file if the strip program failed for any
reason. Before, that file was left behind, sometimes even with wrong
permissions.
[This bug was present in "the beginning".]
ln --relative now updates existing symlinks correctly. Previously it based
the relative link on the dereferenced path of an existing link.
[This bug was introduced when --relative was added in coreutils-8.16.]
ls --recursive will no longer exit with "serious" exit code (2), if there
is an error reading a directory not specified on the command line.
[Bug introduced in coreutils-5.3.0]
mkdir, mkfifo, and mknod now work better when creating a file in a directory
with a default ACL whose umask disagrees with the process's umask, on a
system such as GNU/Linux where directory ACL umasks override process umasks.
[bug introduced in coreutils-6.0]
mv will now replace empty directories in the destination with directories
from the source, when copying across file systems.
[This bug was present in "the beginning".]
od -wN with N larger than 64K on a system with 32-bit size_t would
print approximately 2*N bytes of extraneous padding.
[Bug introduced in coreutils-7.0]
rm -I now prompts for confirmation before removing a write protected file.
[Bug introduced in coreutils-6.8]
shred once again uses direct I/O on systems requiring aligned buffers.
Also direct I/O failures for odd sized writes at end of file are now handled.
[The "last write" bug was introduced in coreutils-5.3.0 but masked
by the alignment bug introduced in coreutils-6.0]
tail --retry -f now waits for the files specified to appear. Before, tail
would immediately exit when such a file is initially inaccessible.
[This bug was introduced when inotify support was added in coreutils-7.5]
tail -F has improved handling of symlinks. Previously tail didn't respond
to the symlink target (re)appearing after being (re)created.
[This bug was introduced when inotify support was added in coreutils-7.5]
** New features
cp, install, mkdir, mknod, mkfifo and mv now support "restorecon"
functionality through the -Z option, to set the SELinux context
appropriate for the new item location in the file system.
csplit accepts a new option: --suppressed-matched, to elide the lines
used to identify the split points.
df --output now accepts a 'file' field, to propagate a specified
command line argument through to the output.
du accepts a new option: --inodes to show the number of inodes instead
of the blocks used.
id accepts a new option: --zero (-z) to delimit the output entries by
a NUL instead of a white space character.
id and ls with -Z report the SMACK security context where available.
mkdir, mkfifo and mknod with -Z set the SMACK context where available.
id can now lookup by user ID, in addition to the existing name lookup.
join accepts a new option: --zero-terminated (-z). As with the sort,uniq
option of the same name, this makes join consume and produce NUL-terminated
lines rather than newline-terminated lines.
uniq accepts a new option: --group to print all items, while separating
unique groups with empty lines.
shred accepts new parameters to the --remove option to give greater
control over that operation, which can greatly reduce sync overhead.
shuf accepts a new option: --repeat (-r), which can repeat items in
the output.
** Changes in behavior
cp --link now dereferences a symbolic link as source before creating the
hard link in the destination unless the -P,--no-deref option is specified.
Previously, it would create a hard link of the symbolic link, even when
the dereferencing options -L or -H were specified.
cp, install, mkdir, mknod and mkfifo no longer accept an argument to the
short -Z option. The --context equivalent still takes an optional argument.
dd status=none now suppresses all non fatal diagnostic messages,
not just the transfer counts.
df no longer accepts the long-obsolescent --megabytes option.
stdbuf now requires at least one buffering mode option to be specified,
as per the documented interface.
** Improvements
base64 encoding throughput for bulk data is increased by about 60%.
md5sum can use libcrypto hash routines where allowed to potentially
get better performance through using more system specific logic.
sha1sum for example has improved throughput by 40% on an i3-2310M.
This also affects sha1sum, sha224sum, sha256sum, sha384sum and sha512sum.
stat and tail work better with EFIVARFS, EXOFS, F2FS, HOSTFS, SMACKFS, SNFS
and UBIFS. stat -f --format=%T now reports the file system type, and tail -f
now uses inotify for files on all those except SNFS, rather than the default
(for unknown file system types) of issuing a warning and reverting to polling.
shuf outputs subsets of large inputs much more efficiently.
Reservoir sampling is used to limit memory usage based on the number of
outputs, rather than the number of inputs.
shred increases the default write block size from 12KiB to 64KiB
to align with other utilities and reduce the system call overhead.
split --line-bytes=SIZE, now only allocates memory as needed rather
than allocating SIZE bytes at program start.
stty now supports configuring "stick" (mark/space) parity where available.
** Build-related
factor now builds on aarch64 based systems [bug introduced in coreutils-8.20]
3.01a22:
mkisofs includes a new option, -legacy, that allows it to reenable
the short options -H/-L/-P; these have been disabled in 2006 for
compatibility with scripts that have not been updated during the
past 10 years. libsiconv now only tries to open a file when the
argument includes a slash in its name. libschily::printf() now
includes support for %n$ argument reordering for the first 30
arguments.
either because they themselves are not ready or because a
dependency isn't. This is annotated by
PYTHON_VERSIONS_INCOMPATIBLE= 33 # not yet ported as of x.y.z
or
PYTHON_VERSIONS_INCOMPATIBLE= 33 # py-foo, py-bar
respectively, please use the same style for other packages,
and check during updates.
Use versioned_dependencies.mk where applicable.
Use REPLACE_PYTHON instead of handcoded alternatives, where applicable.
Reorder Makefile sections into standard order, where applicable.
Remove PYTHON_VERSIONS_INCLUDE_3X lines since that will be default
with the next commit.
Whitespace cleanups and other nits corrected, where necessary.
Reason for updating: numerous security vulnerabilities have been fixed,
and many changes made - functional and fixes. The list is really too long
to include here:
-rw-r--r-- 1 agc wheel 3437489 Jan 24 18:26 /tmp/news
from the advisory:
Malicious or misbehaving unprivileged guests can cause the host or other
guests to malfunction. This can result in host-wide denial of service.
Privilege escalation, while seeming to be unlikely, cannot be excluded.
Only PV guests can take advantage of this vulnerability.
(CVE-2014-1666)
bump PKGREV
* Add the --files option, to list individual files in the various
reporting modes.
* Flexibly report sizes in Kb, Mb, Gb etc as appropriate. The
previous fixed Mb was inconvenient at both ends. Original patch
from James Beal, though I've polished it pretty much into
unrecognisability.
* Make the existing -d (depth) option apply to the -H (static HTML
report) mode, transforming its output from a single HTML file
giving a report for one directory with no crosslinks to a collection
of HTML files with crosslinks between them.
* Introduce a --cgi mode, to make it easy to plumb agedu's web
reporting into an existing web server as an alternative to running
a dedicated one of its own.
* Switch all the HTML-based reporting modes (the internal httpd, the CGI
mode and the dump of static HTML files) to using URIs and filenames
based on the text of the pathname being reported on, rather than
its numeric index in the data file. The aim is that sub-URIs
should remain valid when the data is updated - if, for instance,
you're running the agedu CGI script permanently and changing the
data file under it every so often.
* Suggestion from James Beal: support a '--title' option to override the
'agedu:' prefix at the start of the title of output web pages.
Important changes since last release:
19Feb13
- Fix build/configure problems with bpluginfo.c
- Refactor lock_volumes so most lock a vol rather than globally
- Add virtualfull-extreme test
- Apply patch for chio-changer-openbsd -- Implements listall
- Add bat Mac patch
- Fix bug cannot run Copy/Migrate jobs from bat
- Fix loading of bat translations.
- Fix text input in bat.
- Remove Device from show command as it is not used or updated
- Fix update enabled keyword in help command
- Fix update_postgresql_tables 10 to 11
- Clean after building doc binaries
- Require Qt version 4.8.4
- Fix bug that OK to run? does not retry on bad response
- Backport new lock calls + debug for SD
- Fix bug new label ignored when first is duplicate.
- Pull src/lib changes from master
- Pull SD files from master
- Turn off prune-test
- Add logdir and bsrdir to regression config
- Add bsrdir and logdir to regress-config
- Changed default directories from /var/bacula to /opt/bacula
- Make spool size max message more explicit
- Display more info when maximum spool size is reached
- Fix TERM defs for Solaris 11 in conio.c
- Allow group to cd to sysconfdir
- Make sysconfdir belong to Bacula user/group and exclude others
- Fix Virtual Full file close bug -- race condition
- Enhance mount message to include read/append
- Improve vol_mgr swap messages
- Enhance output of Using Device to include for read/write
- Add jobid to Pmsg output
- Fix bug input validation on delete of jobs.
- Fix bug Authorization Errors false
- Add workaround for where autoprune and reload are in deadlock
- Allow conf file quoted strings to be used in a list
- Make bfgets handle very long lines
- Avoid seg fault by checking for NULL client -- reported by Arno
- Fix bug MailOnSuccess get executed on error.
- Fix missing index on Media table
- Fix bug no message storage on closed database connection.
- Fix bug Fix problem with MySQL with big Base jobs.
- Reduce wait time in regression RunScripts
- Implement a timeout on regression tests of 15 minutes
- Fix .bvfs_lsfiles and .bvfs_restore to handle deleted files
- Fix spooldata, accurate and ingnoreduplicate run argument
- Fix messages segfault
- fix about PATH_MAX on hurd
- fix about MAXPATHLEN on hurd
- Fix bug OpenBSD autochanger example script
- Fix about MaxRunTime canceling the job too early
- Fix bug Wrong values at media list in BAT
- Fix race condition in close_msg that causes seg fault
- Fix double unlink
- More tests for freeing NULL pointers
- Add MaximumConcurrentReadJobs directive to Storage resource
- Add ujobid to .bvfs_get_jobids
- Avoid problem when stripping a relative path
- Fix about HIDDEN attribute set to parent directory during restore
* support for animated images in view(1) and in the preview plug-in
* further improvements to the preview plug-in (zooming images...)
* improved support for translations
* better support for Gtk+ 3.0 (enabled in wip)
time_t on NetBSD 6+.
The configure patch had to be generated with autoconf downgraded to 2.61 -
leave it to autohell to ruin your day...
While we are here, shake, rattle & roll the patches.
The resulting gyrations finally fix PR pkg/46555
This release adds some workarounds for compiling with MinGW. New
functions xcomerr(), xcomerrno(), fxcomerr(), and fxcomerrno() in
libschily. A bug in libschily/format.c that caused %.*s to be
printed incorrectly has been fixed. libschily/fprformat.c gives
better printf() performance on Solaris. Some bugs in the mkisofs
man page have been fixed. Mkisofs now opens "NUL" on Win-DOS instead
of "/dev/null".
All:
- New Schily Makefiles rules for armv7l-linux
- autoconf/xconfig.h.in now treats MinGW32 the same as when compiling
with cl.exe
Libschily:
- libschily::format.c (printf) now supports length modifiers
like h hh l ll t z for the %n format.
- libschily::format.c has been restructured for parsing the
field width parameters in a way that would allow to later
introduce support for %n$
- Try to take care about the fact that MinGW32 does not support
a standard compliant mkdir() function with two parameters.
This applies top libschily/mkdirat.c and libschily/mkdirs.c
- Corrected a typo in libschily/mkfifoat.c that called mkdir()
instead of mkfifo()
- libschily/getdtablesize.c needs the same treatment for MinGW32 as for
cl.exe
Libparanoia (Ported/enhanced by Jörg Schilling, originated by Monty xiphmont@mit.edu):
- libparanoia now implements the first stage in C2 error pointer
support: it now is able to deal with input data that contains
C2 pointers and it implements new callbacks to report C2
statistics back to the caller.
- libparanoia now supports to control the size of the read ahead buffer.
This allows libparanoia to be adopted to the current constraints.
This release includes a veritable ton of work, but the biggest change
is probably the addition of real (YMMV) metadata support, including
hardlinks, and the ability to directly save/restore trees without tar
and split/join. See bup-index(1), bup-save(1), bup-restore(1),
bup-meta(1), bup-xstat(1), and the updated "Using bup" section in the
README.
Note though, that the metadata support really is our first pass, and
there are already things that we know need fixing (i.e. better support
for cross-filesystem-type save/restore (too noisy), etc.). And check
the "Notes on ..." sections in the README for some platform-specific
limitations.
An incomplete list of other notable changes since 0.24b:
- bup will no longer create ~/.bup implicitly; "bup init" is
required.
- "bup split" now supports a compression-level option (-#).
- "bup tag" now supports "-f".
- "bup ls" now supports "-a", "-s", "--human-readable", etc., and
reports more information.
- "bup web" now supports "--human-readable".
- "bup import-rdiff-backup" has been added.
- "bup cat-file" has been added.
- The default "/usr" installation prefix can be overridden via PREFIX.
- Python 2.4 is no longer supported.
And we have at least one (likely minor) known issue:
- *Very* large (i.e. probably greater than MAX_LONG), or negative
filesystem timestamps may not always be handled correctly at the
moment. We'll fix that soon. And related -- some of the tests
may fail on FUSE filesystems.
