proxies slow to update??

[js1]

I'm running Zabbix server and proxy 1.6.4. I just changed the Host name (in the GUI) of a few system I'm monitoring with proxies using external checks. For some, it didn't miss a beat: name changed, no big deal, keep gathering data. For a few others proxies, it missed maybe a minute or two of checks. But, for another set of proxies, I had to restart the the proxy process before the checks resumed. Has anyone else experienced something like this? Other than for active checks, the name of a host configured in the GUI is really just an abstract identifier. The DNS name and IP address attributes of these hosts I updated didn't change. It seems like the checks themselves should not have been affected at all.

I did turn on debugging on the proxies I had to restart, and it seems like on restart it was able to update its database and resume checking:

 32124:20091021:023319 Received [{
        "drules":{
                "fields":["druleid","iprange","delay","status"],
                "data":[]},
        "dchecks":{
                "fields":["dcheckid","druleid","type","key_","snmp_community","ports"],
                "data":[]},
        "hosts":{
                "fields":["hostid","host","dns","useip","ip","port","status","useipmi","ipmi_port","ipmi_authtype","ipmi_privilege","ipmi_username","ipmi_password"],
                "data":[
                        [10224,"New_Name_for_dns1","dns1.mydomain.bogus",1,"10.10.1.1",10050,0,0,623,-1,2,"",""],
                        [10225,"New_Name_for_dns2","dns2.mydomain.bogus",1,"10.10.2.1",10050,0,0,623,-1,2,"",""],
        "items":{
                "fields":["itemid","type","snmp_community","snmp_oid","snmp_port","hostid","key_","delay","status","value_type","trapper_hosts","units","multiplier","delta","snmpv3_securityname","snmpv3_securitylevel","snmpv3_authpassphrase","snmpv3_privpassphrase","formula","logtimefmt","templateid","valuemapid","delay_flex","params",
"ipmi_sensor"],
                "data":[
                        [38417,10,"public","interfaces.ifTable.ifEntry.ifInOctets.1",161,10224,"getdnsresponsetime.sh[]",20,0,0,"","",0,0,"",0,"","","1","",24412,0,"","DSN=<database source name>\\nuser=<user name>\\npassword=<password>\\nsql=<query>",""],
                        [38418,10,"public","interfaces.ifTable.ifEntry.ifInOctets.1",161,10225,"getdnsresponsetime.sh[]",20,0,0,"","",0,0,"",0,"","","1","",24412,0,"","DSN=<database source name>\\nuser=<user name>\\npassword=<password>\\nsql=<query>",""],

 32124:20091021:023319 Query [update hosts set host='New_Name_for_dns1',dns='dns1.mydomain.bogus',useip=1,ip='10.10.1.1',port=10050,status=0,useipmi=0,ipmi_port=623,ipmi_authtype=-1,ipmi_privilege=2,ipmi_username='',ipmi_password='' where hostid=10224;
update hosts set host='New_Name_for_dns2',dns='dns2.mydomain.bogus',useip=1,ip='10.10.2.1',port=10050,status=0,useipmi=0,ipmi_port=623,ipmi_authtype=-1,ipmi_privilege=2,ipmi_username='',ipmi_password='' where hostid=10225;

Why did some proxies not miss a beat while others got delayed while even others had to be restarted?

[richlv]

what is ConfigFrequency set for those proxies in the config files ? did you wait for at least that time to see whether configuration changes propagate ?

[js1]

what is ConfigFrequency set for those proxies in the config files ? did you wait for at least that time to see whether configuration changes propagate ?

It's set for the default 3600 seconds. And, yes, I did wait over an hour. What doesn't make sense is even if it doesn't get the latest configs, the proxies should have continued with the old config info. So, the ConfigFrequency seems like it shouldn't matter in this instance.

[richlv]

were there any relevant messages in the logfile at the time when proxies stopped sending in the data ?

[js1]

were there any relevant messages in the logfile at the time when proxies stopped sending in the data ?

Just combed through the logs and I didn't find any relevant messages. If I had to guess, the proxies were probably operating correctly as they would not have known to do otherwise. Maybe it's the way the Zabbix server handled the data after the name change. For this particular check, I'm keeping 1 day history and 365 day trend. And, in the trend graph for the day I changed the host names, I don't see the gap in data that I was seeing the day of the change. I'm not sure if the trend data has just smoothed out the gap or whether the data was always there and just wasn't properly displayed.