DM 1.6-stable configuration sync problem

[stalker]

How-to initiate full-sync between master and slave node?

I have problem: after adding some items and triggers I see it only on master node. When i on next day connect directly to slave node i do not see this items and triggers, but on master node it still present. After export slave config from master node and import it directly on slave node I recieve trigger notifications from master node with "*UNKNOWN*" hostname in trigger. Both servers alredy restarted. How-to initiate full-sync between master and slave node?

[vinny]

With 1.6 version, the zabbix processes are stronger, i.e they crash less on sync failures...
So u could have sync failure, only visible in the logs.

Do ur master tries to sync with always the same amount of data, or more & more ?


vinny

[stalker]

when i restart slave node i see:

[long sql code with many union]
...
union all select 2,'timeperiods',timeperiodid,1,concat_ws(',',timeperiod_type,every,month,dayofweek,day,start_time,period,date) from timeperiods where 1=1 and timeperiodid between 200000000000000 and 299999999999999
union all select 2,'regexps',regexpid,1,concat_ws(',',md5(name),md5(test_string)) from regexps where 1=1 and regexpid between 200000000000000 and 299999999999999
union all select 2,'expressions',expressionid,1,concat_ws(',',regexpid,md5(expression),expression_type,md5(exp_delimiter),case_sensitive) from expressions where 1=1 and expressionid between 200000000000000 and 299999999999999
] Deadlock found when trying to get lock; try restarting transaction [1213]

and after this i see normal notification sending.

I make this: export slave config from master, delete hosts on both nodes and import config from master node. After 15 minutes configuration has been synced, but notifications not sended Next, after 12 hours, notifications still not send and i restarted slave node. Then it show above error, synced config and all alerts on this time period and after 30 minutes i have duplicated hosts with worked duplicated alarms. Next I delete on master node unneeded duplicates and now it seems functionally.