[ZBX-9769] ZBX_TCP_READ() from Agent on Centos 6.6 Created: 2015 Aug 10  Updated: 2017 May 30  Resolved: 2015 Aug 10

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Agent (G)
Affects Version/s: 2.2.7, 2.4.5
Fix Version/s: None

Type: Incident report Priority: Minor
Reporter: MightyDok Assignee: Unassigned
Resolution: Won't fix Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Centos 6.6, Vmware Cloud, Hosting Provider RTCOMM


 Description   

After install zabbix agent 2.2 from epel and add servers to zabbis server i catch errors in log file: "failed: another network error, wait for 15 seconds".
After turn on debug log errors changed to: "ZBX_TCP_READ() failed: [4] Interrupted system call", and from network dump i see that agent hangup and dont send data more than 20 sec for each request.
Upgrade to zabbix agent 2.4.5 dont solve the problem.

After change Server= parameter in agent conf file error is gone.

Hosting provider DNS server dont have any rate limit, so i think there is some bug in agent.

 Comments   
Comment by MightyDok [ 2015 Aug 10 ]

After change Server= parameter in agent conf file to IP address - error is gone.

Comment by richlv [ 2015 Aug 10 ]

that sounds like misconfiguration of the "Server" parameter then - could you please clarify where is the bug in zabbix ?

edit : or did you mean that changing "Server" parameter from a dns entry to an ip address avoids this problem ?
if so, that's still no indication of any problem in zabbix

Comment by MightyDok [ 2015 Aug 10 ]

@richlv - yep, zabbix server available by FQDN and when i use it for Server parameter then all goes wrong.
But when i change FQDN to IP, or edit /etc/hosts - then all works fine.

Also when i use FQDN, then zabbix server poller and unreachable poller busy status rised to 70-80%. Even when i set timeout to 20 and rise poller process number.

On link an example of network traffic, http://joxi.ru/RmzQPL3tvPL6rO
As you see zabbix agent hang up with simple check: system.cpu.load[percpu,avg1], and there is no any network or load problem with VM. So this problem only with zabbix agent, DNS servers works well.

5666 - custom zabbix port, and 117 IP - address of zabbix server.

Comment by richlv [ 2015 Aug 10 ]

there still does not seem to be any indication of a bug in zabbix, please see https://www.zabbix.org/wiki/Getting_help for support options

Comment by MightyDok [ 2015 Aug 10 ]

Ok, nice. No debug log request, or strace, nothing. Just GTFO. Great support. Thnx.

Generated at Fri Jul 04 10:13:12 EEST 2025 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.