Announcement

Collapse
No announcement yet.

unable to update '/usr/local/oreon/rrd/1.rrd'

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • unable to update '/usr/local/oreon/rrd/1.rrd'

    I have configured 10 hosts with graph_ping, and the commands sometimes run well, but other times gives:

    Code:
    unable to update '/usr/local/oreon/rrd/1.rrd' : illegal attempt to update using time 1162580943 when last update time is 1162580943 (minimum one second step)
    Even running it from the command line gives that. Do i have to configure a diferent rrd file for each host or something? nagios.og, shows that on every host. I guess it's ok that all hosts writes to the same global rrd file, or not? I am very confused

  • #2
    In all case you need different rrd for every check

    Maybe your server need sntp sync
    Intel(R) Xeon(TM) CPU 3.4GHz - MemTotal : 1034476 kB
    Centreon 2.4.1 - Nagios 3.2.1 - Nagios Plugins 1.4.15 - Manubulon Plugins tuné
    Fedora Core 5 - 2.6.20-1.2320

    Comment


    • #3
      Originally posted by DonKiShoot
      In all case you need different rrd for every check
      Humm, i thought oreon did this automatically. Should i do it manually? How? But, i'm seeing there is already many files on the rrd dir, so must be something else.

      Maybe your server need sntp sync
      I run ntpdate from a daily cron script, should i run ntpd instead? It guess ntpd provides a "smoother" time correction. This machine only has 256 MB ram, so i am trying to cut on the daemons running.

      Comment


      • #4
        Re: unable to update '/usr/local/oreon/rrd/1.rrd'

        Originally posted by sansancasd
        I guess it's ok that all hosts writes to the same global rrd file, or not? I am very confused
        It sound like. How did you configure your check_graph_ping ? command and arguments ?
        StatusMap Module - NDO Tools Module - ImportCSV Module - SNMP-UI Module - PDFReports Module
        Dons Paypal

        Comment


        • #5
          It's the default with some minor changes:

          $USER1$/check_graph_ping.pl -H $HOSTADDRESS$ -w 200,20% -c 500,40% -f -n 5 -g -S $ARG1$


          I changed from "-n $ARG1$" to "-n 5" and "-S $ARG2$" to "-S $ARG1$"

          Comment


          • #6
            you don't need to specify argument for -S. Try this one :

            $USER1$/check_graph_ping.pl -H $HOSTADDRESS$ -w 200,20% -c 500,40% -f -n 5 -g -S
            without any argument
            StatusMap Module - NDO Tools Module - ImportCSV Module - SNMP-UI Module - PDFReports Module
            Dons Paypal

            Comment


            • #7
              ok, i deleted the $ARG1$ from -S in graph_ping and graph_nt, let's see how it goes.


              oh, and BTW, many default commands have $ARG1$ on it, so it can't be that wrong.

              Comment


              • #8
                wistof, i tried what you said, but it's the same thing:

                Code:
                unable to update '/usr/local/oreon/rrd/1_1.rrd' : illegal attempt to update using time 1162857558 when last update time is 1162857558 (minimum one second step)
                The curious thing, is that it only complains about 1_1.rrd and 1.rrd when the rrd dir has 94 files! BTW, the logs show that the 1.rrd file is being for a ping check and the 1_1.rrd for a CPULOAD query with check_graph_nt. The ping and CPULOAD check commands were the ones where i took off the $ARG1$.

                Comment


                • #9
                  Make good check_command

                  Suppress service who are not ok.
                  Delete their rrd file.

                  Restart all

                  Recreate services with the good check_command

                  Restart all

                  Oreon should give good name to the rrd file and no more conflict should append
                  Intel(R) Xeon(TM) CPU 3.4GHz - MemTotal : 1034476 kB
                  Centreon 2.4.1 - Nagios 3.2.1 - Nagios Plugins 1.4.15 - Manubulon Plugins tuné
                  Fedora Core 5 - 2.6.20-1.2320

                  Comment


                  • #10
                    ok, i did this and so far, it's going well.

                    Thanks DonkiShoot.

                    Comment


                    • #11
                      Well, i'm reaching the confirmation that sometimes nagios doesn't stop correctly and various nagios processes run at the same time. Yes, you say that nagios forks it self, so how come i still see nagios processes after doing a /etc/init.d/nagios stop? And after doing some nagios restarts via oreon i see nagios 6 process when the usual are 2? Sometimes the nagios process is killed correctly, BUT some plugin process are still runing and are left stale. A check_graph_ping process just went this way. I think that's why i got these:

                      "unable to update '/usr/local/oreon/rrd/35_4.rrd' : illegal attempt to update using time 1163519813 when last update time is 1163536230 (minimum one second step)"

                      Yes, doing a rm /usr/local/oreon/rrd/* solves it, but on a production server, that's losing data. Yes, it's only graph's data, but it can be important to compare stuff from days/months ago to now, or other kind of operations.


                      IMHO, something is wrong with nagios init.d script. In fact in the script, you can read a warning about nagios not beeing stoped correctly, there's a fix there, but it doesn't seem to work 100%. Or my setup is messed up

                      Am i the only one seeing this processes mess?

                      Comment


                      • #12
                        I am facing the same issue. I have allready tried deleting the rrd files in /usr/local/oreon/rrd/ directory and restarting, but it did not work. Could there be any other solution?

                        Thanks in advance

                        Comment

                        Working...
                        X