Re: [Nagios-users] FW: NDO Utils Question
-Original Message- From: Christopher McAtackney [mailto:crist...@gmail.com] Sent: 25 June 2009 16:33 Results passing into the command pipe are stored. The relevant parameters are in nagios.cfg; they are external_command_buffer_slots and check_result_buffer_slots - by default these are set to 4096 (see documentation within the configuration file). Great, that's what I was hoping for. Do you have any experience of setting this buffer to much higher values? Not that I necessarily intend to, but it's always useful to know the effects of pushing the system to its limits. Routinely we have external_command_buffer_slots set to 40960 and check_result_buffer_slots set to 61440; this is because we have had problems with our SQL server (it gets very busy for other databases) which delays response to NDOUtils updates which fills up these buffers. You can see the current, high-water mark and setting for these parameters by running command nagiostats (the highest we have reached recently was about 25k buffer slots used). Jonathan Wheeler e-Science Centre Rutherford Appleton Laboratory -- Scanned by iCritical. -- ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
[Nagios-users] FW: NDO Utils Question
-Original Message- From: Christopher McAtackney [mailto:crist...@gmail.com] Sent: 25 June 2009 14:51 Hi everyone, I have a quick question about Nagios and NDOUtils I was hoping someone could answer. What happens if the database that NDO Utils is using becomes unavailable? (e.g. the server has crashed). The answer depends a lot on whether you are using Nagios v2 or Nagios v3; we are using Nagios v2.11. My understanding is that in Nagios v2, the code that communicates with the event broker module in single-threaded. Therefore a problem with the SQL server can jam up Nagios to the extent that it effectively stops running commands. In Nagios v3, threading has been rewritten and this problem no longer exists. I'm assuming Nagios will continue to monitor and send notifications as normal, is this correct? In Nagios v2 it seems that almost all activity is suspended when you are using NDOutils and the MySQL server is unavailable; this continues until the SQL server is restored. One solution is to restart Nagios without the broker_module. What about the service check results that would normally be passed to NDO Utils and then stored in the database? Are they queued somewhere? And if so, how is the capacity of this queue defined? If they are not queued, what happens? Will NDO Utils just throw an error for each result it tries to store in the database and fails? Will this affect the core Nagios process? Results passing into the command pipe are stored. The relevant parameters are in nagios.cfg; they are external_command_buffer_slots and check_result_buffer_slots - by default these are set to 4096 (see documentation within the configuration file). Hopefully someone can provide some insight. Hopefully this has Jonathan Wheeler e-Science Centre Rutherford Appleton Laboratory -- Scanned by iCritical. -- ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
Re: [Nagios-users] FW: NDO Utils Question
2009/6/25 Wheeler, JF (Jonathan) jonathan.whee...@stfc.ac.uk: -Original Message- From: Christopher McAtackney [mailto:crist...@gmail.com] Sent: 25 June 2009 14:51 Hi everyone, I have a quick question about Nagios and NDOUtils I was hoping someone could answer. What happens if the database that NDO Utils is using becomes unavailable? (e.g. the server has crashed). The answer depends a lot on whether you are using Nagios v2 or Nagios v3; we are using Nagios v2.11. My understanding is that in Nagios v2, the code that communicates with the event broker module in single-threaded. Therefore a problem with the SQL server can jam up Nagios to the extent that it effectively stops running commands. In Nagios v3, threading has been rewritten and this problem no longer exists. Sorry, should have specified that. I'm running 3.0.6. I'm assuming Nagios will continue to monitor and send notifications as normal, is this correct? In Nagios v2 it seems that almost all activity is suspended when you are using NDOutils and the MySQL server is unavailable; this continues until the SQL server is restored. One solution is to restart Nagios without the broker_module. What about the service check results that would normally be passed to NDO Utils and then stored in the database? Are they queued somewhere? And if so, how is the capacity of this queue defined? If they are not queued, what happens? Will NDO Utils just throw an error for each result it tries to store in the database and fails? Will this affect the core Nagios process? Results passing into the command pipe are stored. The relevant parameters are in nagios.cfg; they are external_command_buffer_slots and check_result_buffer_slots - by default these are set to 4096 (see documentation within the configuration file). Great, that's what I was hoping for. Do you have any experience of setting this buffer to much higher values? Not that I necessarily intend to, but it's always useful to know the effects of pushing the system to its limits. Hopefully someone can provide some insight. Hopefully this has Yes, this has been very helpful, thanks for your answer. Thanks as well Carsten, much appreciated. Jonathan Wheeler e-Science Centre Rutherford Appleton Laboratory -- ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null