Module: check_mk Branch: master Commit: 80d46419c960b41ba994b74255e9c7645a9ae532 URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=80d46419c960b4… Author: Andreas Boesl <ab(a)mathias-kettner.de> Date: Fri Jul 17 11:00:06 2015 +0200 #2407 FIX WATO master/slave replication: fixed problem where the configuration from the master site was not activate on slave site Changes made on a distributed WATO slave site could delay a configuration update attempt from the master. Detailed information: Any changes on a slave site which lead to a pending log entry triggered the cmcrushd. The cmcrushd creates a configuration for the microcore, which takes precendence over creating the configuration anew. When a snapshot got pushed from the master site, the slave site thought that there is already a precompiled configuration available and used the cmcrushd configuration. This problem is automatically resolved on the following configuration push from the master, because the cmcrushd configuration is only used once. --- .werks/2407 | 18 ++++++++++++++++++ ChangeLog | 1 + web/htdocs/wato.py | 7 ++++++- 3 files changed, 25 insertions(+), 1 deletion(-) diff --git a/.werks/2407 b/.werks/2407 new file mode 100644 index 0000000..00d3674 --- /dev/null +++ b/.werks/2407 @@ -0,0 +1,18 @@ +Title: WATO master/slave replication: fixed problem where the configuration from the master site was not activate on slave site +Level: 2 +Component: wato +Class: fix +Compatible: compat +State: unknown +Version: 1.2.7i3 +Date: 1437116656 + +Changes made on a distributed WATO slave site could delay a configuration update attempt from the master. + +Detailed information: Any changes on a slave site which lead to a pending log entry triggered the cmcrushd. +The cmcrushd creates a configuration for the microcore, which takes precendence over creating the configuration anew. +When a snapshot got pushed from the master site, the slave site thought that there is already a precompiled configuration +available and used the cmcrushd configuration. + +This problem is automatically resolved on the following configuration push from the master, +because the cmcrushd configuration is only used once. diff --git a/ChangeLog b/ChangeLog index bf7dd49..4f4bed8 100644 --- a/ChangeLog +++ b/ChangeLog @@ -95,6 +95,7 @@ * 2464 FIX: UDP ports for SNMP communication can now be configured via WATO * 2466 FIX: Fixed exception when searching for rulesets / global settings using special characters like umlauts * 2467 FIX: Fixed encoding exception occuring in localized GUI when a WATO action triggers an error message... + * 2407 FIX: WATO master/slave replication: fixed problem where the configuration from the master site was not activate on slave site... Notifications: * 2313 FIX: notification bulking: fixed exception for plugins which are not configured with checkboxes, e.g. sms diff --git a/web/htdocs/wato.py b/web/htdocs/wato.py index a202509..02e0142 100644 --- a/web/htdocs/wato.py +++ b/web/htdocs/wato.py @@ -5119,7 +5119,9 @@ def log_pending(status, linkinfo, what, message, user_id = None): if 'need_to_bake_agents' in globals(): need_to_bake_agents() - if not is_distributed(): + # The latter one condition applies to slave sites + # Otherwise slave sites would trigger the cmcrushd + if not is_distributed() and not has_distributed_wato_file(): if status != SYNC: log_entry(linkinfo, what, message, "pending.log", user_id) cmc_rush_ahead() @@ -11999,6 +12001,9 @@ def delete_distributed_wato_file(): if os.path.exists(p): create_user_file(p, "w").write("") +def has_distributed_wato_file(): + return os.path.exists(defaults.check_mk_configdir + "/distributed_wato.mk") + #. # .--Users/Contacts------------------------------------------------------. # | _ _ ______ _ _ |