Ad Widget

Collapse

Zabbix sender messy output

Collapse
X
Collapse
 
  • Page of 1
  • Time
  • Show
Clear All
new posts
Previous template Next
  • jsosic
    Member
    • Apr 2008
    • 47
    #1

    Zabbix sender messy output

    Hi,

    I've observed some scewed behaviour of zabbix_sender in versions 2.0+.

    In previous version (1.8), I could normaly pipe output of zabbix_sender, but not after upgrade to 2.0.

    Here are the examples:

    Code:
    # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14
info from server: "Processed 0 Failed 1 Total 1 Seconds spent 0.000017"
sent: 1; skipped: 0; total: 1
    Code:
    # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | head -1
sent: 1; skipped: 0; total: 1
    And again:

    Code:
    # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | grep -v pero
sent: 1; skipped: 0; total: 1

# zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | sed 's/foo/bar/'
sent: 1; skipped: 0; total: 1

# zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | tee
sent: 1; skipped: 0; total: 1
    And now piping:
    Code:
    # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 2>&1 | tee
sent: 1; skipped: 0; total: 1

    OK, so what's up with this? strace shows the following:

    Code:
    ...
...
munmap(0x7f3079add000, 4096)            = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7f3079ad49d0) = 13637
wait4(13637, info from server: "Processed 0 Failed 1 Total 1 Seconds spent 0.000023"
[{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 13637
--- SIGCHLD (Child exited) @ 0 (0) ---
fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 0), ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f3079add000
write(1, "sent: 1; skipped: 0; total: 1\n", 30sent: 1; skipped: 0; total: 1
) = 30
exit_group(0)                           = ?

    Is this a bug, or what?!
    Tags: None
    • neogan
      Senior Member
      Zabbix Certified Trainer
      Zabbix Certified SpecialistZabbix Certified Professional
      • Sep 2011
      • 118
      #2
      Originally posted by jsosic
      Hi,

      I've observed some scewed behaviour of zabbix_sender in versions 2.0+.

      In previous version (1.8), I could normaly pipe output of zabbix_sender, but not after upgrade to 2.0.

      Here are the examples:

      Code:
      # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14
info from server: "Processed 0 Failed 1 Total 1 Seconds spent 0.000017"
sent: 1; skipped: 0; total: 1
      Code:
      # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | head -1
sent: 1; skipped: 0; total: 1
      And again:

      Code:
      # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | grep -v pero
sent: 1; skipped: 0; total: 1

# zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | sed 's/foo/bar/'
sent: 1; skipped: 0; total: 1

# zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 | tee
sent: 1; skipped: 0; total: 1
      And now piping:
      Code:
      # zabbix_sender -c zabbix_agentd.conf -k mmysql.QCInserts -o 14 2>&1 | tee
sent: 1; skipped: 0; total: 1

      OK, so what's up with this? strace shows the following:

      Code:
      ...
...
munmap(0x7f3079add000, 4096)            = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7f3079ad49d0) = 13637
wait4(13637, info from server: "Processed 0 Failed 1 Total 1 Seconds spent 0.000023"
[{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 13637
--- SIGCHLD (Child exited) @ 0 (0) ---
fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 0), ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f3079add000
write(1, "sent: 1; skipped: 0; total: 1\n", 30sent: 1; skipped: 0; total: 1
) = 30
exit_group(0)                           = ?

      Is this a bug, or what?!
      Registred already ZBX-6285.
      but another environment.

        Comment

        • vintagegamingsystems
          Member
          • Jun 2013
          • 57
          #3
          Efficient logging of zabbix_sender command is impossible

          There doesn't seem to be any way to pipe or redirect the command without standard output. I can't efficiently log the output of the command due to this problem. I may look at the code and see what is wrong with this.

            Comment

            • steveboyson
              Senior Member
              • Jul 2013
              • 582
              #4
              I noticed this as well.
              In our perl scripts where we rely on the FULL output of zabbix-sender, we are using IPC::Run.

              Found here:
              http://stackoverflow.com/questions/1...perl-open-pipe

                Comment

                Previous template Next
                Working...