>> Here are the results, i repeat that i only appears using poller2. > Hmm, thats strange in itself, but let's see.
I think you are right, I notice it with poller2 only because I have a huge number of hosts now and the problem only appears when the interface flaps (UP/DOWN) within the poll cycle, when the interface stay down the duration is OK. > OK, here is our problem. Can you run > SELECT *, (date_stop - date_start) AS duration FROM alarms WHERE triggered = > 1 AND interface=3108; For sure, the duration calculation into mysql is bad mysql> SELECT *, (date_stop) - (date_start) AS duration FROM alarms WHERE triggered = 1 AND interface=3108; +------+---------------------+---------------------+-----------+------+--------+---------------+--------------+-----------+-------------+ | id | date_start | date_stop | interface | type | active | referer_start | referer_stop | triggered | duration | +------+---------------------+---------------------+-----------+------+--------+---------------+--------------+-----------+-------------+ | 1598 | 2007-10-04 21:57:56 | 2007-10-04 22:06:20 | 3108 | 40 | 12 | 2147 | 2162 | 1 | 4864.000000 | | 1652 | 2007-10-04 23:01:18 | 2007-10-04 23:09:22 | 3108 | 40 | 12 | 2220 | 2223 | 1 | 804.000000 | ---snip----- 23 rows in set (0.00 sec) mysql> SELECT *, TIME_TO_SEC(TIMEDIFF((date_stop),(date_start))) AS duration FROM alarms WHERE triggered = 1 AND interface=3108; +------+---------------------+---------------------+-----------+------+--------+---------------+--------------+-----------+----------+ | id | date_start | date_stop | interface | type | active | referer_start | referer_stop | triggered | duration | +------+---------------------+---------------------+-----------+------+--------+---------------+--------------+-----------+----------+ | 1598 | 2007-10-04 21:57:56 | 2007-10-04 22:06:20 | 3108 | 40 | 12 | 2147 | 2162 | 1 | 504 | | 1652 | 2007-10-04 23:01:18 | 2007-10-04 23:09:22 | 3108 | 40 | 12 | 2220 | 2223 | 1 | 484 | ---snip--- 23 rows in set (0.00 sec) I attached the patch for the event api. Other thing, I don't know the impact of having time between 2 poll cycles > 5 minutes (for RRD ?) I tried to increase the number of poller threads but I notice that with thread > 40 the system is overloaded (IO > 15 Mb/s and cpu increasing to 100% and load average jumping to 60), with poller thread = 39 I have a correct value for loadaverage, CPU, and IO rate) but my poll cycle is > 5 Minutes ... I plan to add up to 10 000 interface on this system. Migrating the database to another box does not solve the overload problem, I don't know if I'm facing a limit, a misconfiguration or a BUG somewhere. Any advice will be greatly appreciated. _______________ David LIMA Professional Services www.scc.com -----Message d'origine----- De : [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] De la part de Craig Small Envoyé : samedi 6 octobre 2007 00:07 À : jffnms-users@lists.sourceforge.net Objet : Re: [jffnms-users] Consolidate and poller2 bug On Fri, Oct 05, 2007 at 07:42:33PM +0200, LIMA David wrote: > Here are the results, i repeat that i only appears using poller2. Hmm, thats strange in itself, but let's see. > mysql> SELECT (date_stop - date_start) AS duration FROM alarms WHERE > triggered = 1 AND interface=3108; > +-------------+ > | duration | > +-------------+ > | 4864.000000 | OK, here is our problem. Can you run SELECT *, (date_stop - date_start) AS duration FROM alarms WHERE triggered = 1 AND interface=3108; > | 2147 | 2007-10-04 21:57:56 | 40 | 306 | Reachability Test | unreachable > | status | 80% Packet Loss | 0 | 1 | 1 | > | 2162 | 2007-10-04 22:06:20 | 40 | 306 | Reachability Test | reachable | > status | 0% Packet Loss | 0 | 1 | 1 | So they're the right times as before. - Craig -- Craig Small GnuPG:1C1B D893 1418 2AF4 45EE 95CB C76C E5AC 12CA DFA5 http://www.enc.com.au/ csmall at : enc.com.au http://www.debian.org/ Debian GNU/Linux, software should be Free ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ jffnms-users mailing list jffnms-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jffnms-users ______________________________________________________________________ Ce message contient des informations dont le contenu est susceptible d'etre confidentiel. Il est destine au(x) destinataire(s) indique(s) exclusivement. A moins que vous ne fassiez partie de la liste des destinataires, ou que vous soyez habilite a recevoir le mail a leur place, il vous est interdit de le copier, de l'utiliser ou de devoiler son contenu a un tiers. Si vous avez recu cet email par erreur, merci de prendre contact avec l'emetteur. Les opinions exprimees dans cet e-mail sont celles de l'emetteur et ne refletent pas necessairement celles de l'entreprise. Ce e-mail peut contenir des pieces jointes dont certaines pourraient contenir des virus qui pourraient endommager votre systeme informatique. La compagnie a pris toutes dispositions afin de minimiser ce risque et decline toute responsabilite pour toute perte ou dommage resultant directement ou indirectement de l'utilisation de cet email ou de son contenu. Il vous appartient d'effectuer vos propres controles anti-virus avant d'ouvrir la ou les pieces jointes. ______________________________________________________________________
duration_patch
Description: duration_patch
------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________ jffnms-users mailing list jffnms-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jffnms-users
