NetXMS Support Forum

Hello!

What is your MySQL version? So far it looks like your MySQL server does not support prepared statements, which is the case for versions before 4.1.

Best regards,
Victor

A chto znachit "cherez IP"? SNMP? Togda UPS nado dobavit' kak otdel'noe SNMP ustrojstvo i nastroit' sootvetstvujuschie parametri. UPS subagent prednaznachen tol'ko dlja monitoringa cherez serial port i USB.

Quote from: Akira on March 29, 2012, 07:42:41 AM
7. AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA!
После смены IP для одного сервера через Java Console, слетел IP сразу на 5 нодах.

A mozno popodrobnee opisat' posledovatel'nost' dejstvij i rezul'tat? I est' li chto-nibud' v logah?

Quote from: Goriem on March 29, 2012, 03:12:49 PM
I ended up solving it via externalparameter, wrote a vb script to parse the value from the weblog and returning it to the node works like a charm allthough i had to restart the Netxms core server before i could "rearm" the value when falling below configured threshold.

Can you show me configuration of your thresholds? That should not require server restart.

Best regards,
Victor

That's strange. Can you show me results of SNMP walk for your device? For .1.3.6.1.4.1.14685.1.3 for example?

Best regards,
Victor

Quote from: rikm1 on March 29, 2012, 10:08:21 AM
I would strongly recommend Android. The app can be used on tablets and mobiles.

Actually it's already available: https://www.netxms.org/download/netxms-console-1.1.10.R6.apk

Best regards,
Victor

Looks like you don't provide correct instance numbers. I suggest to open MIB Explorer, go to element of interest, and do walk on it. Than copy correct OIDs from result window.

Best regards,
Victor

The consequences for this is that collected data is not stored for node with id 117. You can try to stop NetXMS server, connect to your MySQL database, and run

REPAIR TABLE idata_117;

Then I suggest to run

nxdbmgr check

to check database for consistency. Then you can start NetXMS server again.

Best regards,
Victor

Da, pozaluj smisl v etom est'. Po idee v bol'shinstve situacij propustit' odin poll luchshe chem poluchit' pik na grafike. Ja dumaju sdelat' tak:

1. Dobavit' flazok "is counter" v DCI - poskol'ku eto mozet bit' naprimer temperatura, i plavat' vverh-vniz - togda sled. znacjenie men'she prediduschego eto ok;
2. Esli new value < prev value, to libo propustit' poll (net dannih voobsche), libo sdelat' delta value = prev delta value. V ljubom sluchae pik uberem.

1. Posmotrju popozze.

2-4. Da, eto izvestnie bagi. Budem ispravljat'. #2 dlja menja samij sloznij - nado prosto zakazivat' dizajneru ikonki, poetomu kakoe-to vremja zajmet.

5. Kakie grafiki imejutsja vvidu - predefined? Ili kotorie pokazivajutsja iz last values po right-click?

Vozmozno ne obnovilis' dannie esche. Versija obnovitsja posle configuration poll.

Hi!

Looks like a bug. I was able to reproduce this on my system, so hopefully I'll come out with solution soon.

Best regards,
Victor

That's very strange that cluster nodes does not appear under "entire network". How do you have cluster object configured? Also, what interfaces are seen on cluster nodes?

Best regards,
Victor

это встроенний код, не скрипт. выглядит он так:


      case DCM_AVERAGE_PER_MINUTE:
         nElapsedTime /= 60;  // Convert to minutes
      case DCM_AVERAGE_PER_SECOND:
         // Check elapsed time to prevent divide-by-zero exception
         if (nElapsedTime == 0)
            nElapsedTime++;

         switch(m_dataType)
         {
            case DCI_DT_INT:
               value = ((LONG)value - (LONG)m_prevRawValue) / (LONG)nElapsedTime;
               break;
            case DCI_DT_UINT:
               value = ((DWORD)value - (DWORD)m_prevRawValue) / (DWORD)nElapsedTime;
               break;
            case DCI_DT_INT64:
               value = ((INT64)value - (INT64)m_prevRawValue) / (INT64)nElapsedTime;
               break;
            case DCI_DT_UINT64:
               value = ((QWORD)value - (QWORD)m_prevRawValue) / (QWORD)nElapsedTime;
               break;
            case DCI_DT_FLOAT:
               value = ((double)value - (double)m_prevRawValue) / (double)nElapsedTime;
               break;
            case DCI_DT_STRING:
               // I don't see any meaning in "average delta per second (minute)" for string
               // values, so result will be 0 if there are no difference between
               // current and previous values, and 1 otherwise
               value = (LONG)((_tcscmp((const TCHAR *)value, (const TCHAR *)m_prevRawValue) == 0) ? 0 : 1);
               break;
            default:
               // Delta calculation is not supported for other types
               break;
         }


Quote from: Akira on March 27, 2012, 08:29:36 AM
28-Dec-2011 11:59:01   503850560
28-Dec-2011 10:59:01   795522378
28-Dec-2011 09:59:00   598589222
28-Dec-2011 08:59:00   9223372036854775807
28-Dec-2011 07:59:00   273392862
28-Dec-2011 06:59:00   163655688
28-Dec-2011 05:59:01   43122668
28-Dec-2011 04:59:01   86841056

Я так понимаю, происходит это когда предыдущее значение было больше, чем текущее.

Как можно сделать, чтобы при такой ситуации, просто считалось, что предыдущее значение = "0"?

Само по себе новое значение меньше предыдущего проблем вызывать не должно. Еcли параметр определен как unsigned и размер правильный, то все будет в порядке. Проблема может быть в несовпадении размера данных в DCI и на ноде - если реально данные 32 бита, а DCI определен как 64 бита, то при проходе счетчика через ноль как раз такой эффект и будет.