NetXMS Support Forum

Thank you Victor for a detailed explanation, it's very helpful.

Hi,

Are there any known or recommended values for DCis monitored on the management node? Is any of the following parameter worth watching and having a threshold alert defined?

So far I was able to come up with the following "rules", but I not sure if this is correct. Also some data are missing:
- "Configuration poller queue for last minute" --  depends on server configuration value NumberOfConfigurationPollers
- "Status poller queue for last minute" -- depends on server configuration value NumberOfStatusPollers that should be 1/10th of Nodes.
- "Average time to queue DCI for polling for last minute" -- will increase if DCI polling takes more time - rule should be that (avarage time * number of DCIs) / NumberOfDataCollectors < DCI polling interval
- "Database writer's request queue (other queries) for last minute" -- this and following two are influenced by NumberOfDatabaseWriters server configuration value, but I am not sure when this value needs to be increased.
- "Database writer's request queue (DCI data) for last minute"
- "Database writer's request queue for last minute"
- "Data collection poller's request queue for last minute" -- probably depends on NumberOfDataCollectors and "Average time to queue DCI ..."

Can someone add more info from larger deployments about acceptable values?

Thanks and Regards,
Lukas

"Database Connection Pool exhausted, fallback connection used" is more like a warning that the connection pool has insufficient size and additional connection to database needs to be established. Maybe it is a consequence of some other problem, but as far as I know, this alone should not cause client request timeout.
 
Size of the connection pool is controlled by server configuration variable ConnectionPoolMaxSize, message should disappear if you increase it.
When I leave it at the default size, this message appears about 5 times during the netxms start in my setup.

Hope it helps,
Lukas

Nice.
What can we expect from the feature called "XMPP messaging support"?

Best Regards,
Lukas

Can you provide some context information?
What are those .bit files suppose to contain? Which OS?

Lukas

Not sure if it helps in your case, but you may try to check the following short new howto: http://wiki.netxms.org/wiki/Add_new_MIBs

Lukas

1. http://en.wikipedia.org/wiki/Snmp
2. NetXMS works as SNMP manager for SNMP agents installed on devices - your linux servers probably have SNMP already installed (search for net-snmp, look at /etc/snmp/snmpd.conf)

Hope it helps.
Lukas

Hi,
There is quite flexible way to change the interface expected state via nxshell - see http://wiki.netxms.org/wiki/Using_nxshell_to_automate_bulk_operations#Set_expected_interface_state

Hope it helps,
Lukas

Hi,
seems to me that we are trying to solve the same problem - look at my partial solution, maybe that will fit your needs. https://www.netxms.org/forum/configuration/how-to-monitor-sysuptime-via-snmp/

Regards,
Lukas

Hello,
I am trying to find an optimal way to monitor sysUpTime (OID: .1.3.6.1.2.1.1.3.0) and hrSystemUptime (OID: .1.3.6.1.2.1.25.1.1.0). Both values are timeticks which is not supported type for DCI so I ended up with two following solutions, each with certain drawbacks.

Solution 1 - see the human readable value but no graph
1. Create DCI Internal/Dummy - type String
2. Add transformation script:

transport = CreateSNMPTransport($node);
if (transport == null)
{
  trace(1,"Snmp transport error on ".$node->ipAddr);
  return -1;
}

uptime = SNMPGetValue(transport, ".1.3.6.1.2.1.1.3.0");
if (uptime != null)
{
  // return uptime/100; // in seconds as float
  // return uptime/8640000 // in days as float
  return SecondsToUptime(uptime/100); // readable format
}
else
{
  trace(1,"Could not read uptime value for ".$node->ipAddr);
  return -1;
}


Solution 2 - graph in days, but no easily readable output
Same as above, only change the value type to Floating point and uncomment required line in the script.

Does anybody have a better way for monitoring uptime or Timeticks SNMP values in general so there is both human readable and graphical output?

Regards,
Lukas

Motivation: In larger networks, there can be quite large amount of alarms and events in logs. When netxms user is trying to solve a network problem, he filters existing alarms or events until it is narrow enough to show only required items. Such tasks are typically repeated several times and creating the new filtering expression every time is tedious and error prone. That is why it would be useful if log filters can be saved for later use.

User will need to be able to save and use filter expression, give name to filter expression, change the name and remove filter expression.
I am not quite sure how to fit such feature in existing GUI. One solution from other NMS was, that each unfiltered log (events, alarms, ...) was a node in a tree. Filter expressions saved by user were nodes below that.

What do you think, do you find this feature useful and worth implementing?

Lukas

An article comparing Zenoss Core, Nagios Core, NetXMS and OpenNMS was published at networkworld.com. Access to the article requires a registration.
http://www.networkworld.com/reviews/2013/070813-open-source-network-management-271449.html?page=5

 Hello,
I wonder if anybody has experience with deploying NetXMS as high availability two-nodes solution working over WAN.
DRBD seems to be working fine in the reliable local area network, but generally unreliable WAN links makes it almost impossible to design affordable and reliable solution with automatic start of the standby server and proper split-brain handling. Commercial solutions such as MySQL Cluster or Oracle Extended RAC do support WAN cluster, but it requires either extreme connectivity requirements or at least 4 nodes to operate.

Did anyone had success with one of the commercial solutions or with any other open-source in the similar context?

Kind Regards,
Lukas

Yes, there is a way. Try to change the following script based on your needs: http://wiki.netxms.org/wiki/Using_nxshell_to_automate_bulk_operations#Set_expected_interface_state

Hope it helps,
Lukas

I can confirm that it works now.
Thank you Victor.