Hello,

We see some problems with zabbix server trappers processes. Some processes are just locked with following backtrace
(gdb) bt
#0 0x4c61d402 in __kernel_vsyscall ()
#1 0x4c30b903 in read () from /lib/libc.so.6
#2 0x0807762a in zbx_tcp_recv_ext ()
#3 0x0805c3fa in process_trapper_child ()
#4 0x0805c868 in child_trapper_main ()
#5 0x08055fc2 in MAIN_ZABBIX_ENTRY ()
#6 0x08067499 in daemon_start ()
#7 0x08055abb in main ()
(gdb)
When all trapper processes will be locked in read(), data from senders is stopping to arrive.

After investigating the code, I'm unable to find any timeout for ZBX_TCP_READ/read() from socket or setting socket to non-blocking mode. This read() can hung in a such manner forever.

I see this issue in all 1.4-1.7 verisons.

Thanks,
Alex