Released on February 17, 2010
Improvements
- Added support for Monit 5.1 (required to support multiple service groups per service which was added in Monit 5.1).
- Allow non-fully qualified domain name in mail addresses
- Added automatic synchronization of services and service group names with Monit. Services that are no longer configured in any Monit agent configuration file are not displayed in the rules admin drop-down menu. Likewise, if new services are added to a Monit agent they will be displayed and made available when setting up alert rules in M/Monit. The Events page will still display removed services in the drill-down filter if there are events referring to old removed services.
- When specifying a rule, the service menu is context sensitive, that is, the menu lists only services that belongs to the selected host. If a host group is selected, only services that are common for all hosts in the group.
- M/Monit database schema version added and checked on startup.
Bug fixes
- Fixed alert mail duplication which could occur in some situations. Thanks to George Agnelli for bug report.
- Fixed sporadic long waits when doing an action on a service in the status detailed page. When M/Monit performs a start, stop restart etc action it waits up to 30 seconds for a reply from Monit before it times out. In some situations, especially if Monit is running in a virtual machine, Monit could reply before M/Monit started its wait, which would result in M/Monit waiting the full 30 seconds before timeout.
- Fixed a crash that could occur if a host was added to a non-existent host group. Thanks to Jarod Reid for report.
- Fixed initial sorting of host names in the host admin table.
- If several sessions were open for the same user and password changed, the new password was not applied until all the sessions had terminated. This is now fixed and password changes takes effect immediately.
- Monit 5.1 now makes sure that it sends a timely status message to M/Monit, even if it is busy in long testing cycles, to prevent false alerts generated by M/Monit about non-responsive Monit agents.
Upgrading from M/Monit 2.1.x
If you are upgrading from M/Monit 2.1.1 please perform the following to upgrade the database:
If you use SQLite:
./upgrade/upgrade-2.1.x-2.2.sqlite
If you use MySQL:
cp mmonit-2.1.1/conf/* mmonit-2.2/conf/ cat ./upgrade/upgrade-2.1.x-2.2.mysql | mysql -u mmonit -p -D mmonit
If you use Postgresql:
cp mmonit-2.1.1/conf/* mmonit-2.2/conf/ cat ./upgrade/upgrade-2.1.x-2.2.postgresql | psql -U mmonit mmonit
Browser Cache: Please clear your Browser cache before login to M/Monit 2.2. Some pages have been changed and cached pages from the previous version may create problems.
Known issues
- Mac OS X 10.6.x (Snow Leopard): There is a bug in the system library, libpthread, which, if M/Monit is run in daemon mode will result in that one CPU core will max out and M/Monit becomes unresponsive. The workaround is to only run M/Monit in the foreground using the -i switch to the mmonit program.
- Empty host groups can cause an error in the rule page. If you create host groups with no hosts and restart M/Monit and then go to the admin/rule page a JSON parse error is raised. The workaround until the next release is to delete empty hosts groups or to avoid this issue, make sure at least one host is added to a group.