NetXMS Support Forum

I saw that I can use ServiceResponseTime.HTTP on any agent. But DCI must be under that agent's node. I could not use netsvc. response time yet.

Is it possible to create the ServiceResponseTime.HTTP DCI on service's node and check it with an other node (same  as network service functionality) ? Or is there any other way to do this ?
thanks.

Hi.
I am using 2.0.M4 and have a problem about network service response times. I did not try to use this feature before.
As you can see in attached picture, I have multi web pages in several servers. I created network services for nodes and can see if they are working or not.
I have configured service check as server's agent.

I can load netsvc subagent on server's agent.

But I want to see response times for each network service too. And I could not create a right DCI.
I am creating a DCI in ebysapp1 for ex. and selecting internal, Netsvc.responsetime {instance} but after I apply, nextms is giving me that DCI is unsupported.

I am not writing anything to instance discovery section for new DCI.

Where am I wrong, can anybody explain me?
Thanks.

Hi. I am working with nearly 15k nodes. And had problems while I was using mysql. Now, i am using postgresql and made tuning on it. There is not any problem for me now.

In my opinion, postgresql performance is better then mysql for my work.

if (rindex($1,":")>0)   // this is for your case.
{
// if(length($1)>3) return false;    // in some cases, you can link a disc drive to directory. uncomment this line if you don't want to see them.
if (rindex($1,"A:")>0) return false;  // in my case, i have virtual pcs, and some of them have floppy disk drives, i don't want to see them.
value = AgentReadParameter($node, "FileSystem.Total(" . $1 . ")");
return (value != null) && (value > 0);
//return %(true, $1);
}
return false;

hi.
I am wondering that if dashboards are auto refreshing themselves even if they are in inactive tab of management console.
I mean: I am opening management console, and opening 4-5 dashboards. and one of them is active. Will all of open dashboards refresh in their scheduled times? (in picture below, will DC_STATUS, firewall_Trafik and STATUS dashboards refresh themselves when scheduled interval arrived ?)

if so, can I disable this option. I think this is happening lots of IOPS on my storage, and this is causing problem:
I have many users, they are opening lots of dashboards and at the end, netxms can not write database (when i write show queue in console, I saw 1.500.000 Database writer (IDATA).) Or, why does this happening and what must I do?

thanks.

Btw, I had a look to logs, both mysql and netxmsd, there was 2 error in netxms and 0 in mysql.
errors in netxmsd:  Thread "Item Poller" does not respond to watchdog thread (occurred 2 times in 10 hour)
(added an image)

I think I solved my problem.
I increased Data Collector count in server configuration a few days ago (i have 12000 DCI and am getting value per 10 seconds). But my devices could not answer too much connection at the same time.
So, I decreased it to 50 and now there is no problem.

thank you.

hi.

I am having lots of data collection error, and i could not understand why this is happening.

I added new nodes (routers and servers) to netxms in 10 days. I am  getting this error for 5-6 days.

Anybody has an idea?

never mind. I reinstalled linux, netxms and created a new database.
Deleted all dci value tables from old db and import other tables to new one.
Everything is ok.

No OverCPU usage, No high ram usage no queue.

thank you.

Hello.

I have lots of routers, server pcs and lots of DCI for them.
I have some problems after I upgraded to 2.0-m1.
I am using cents 7, mysql.

server command
show stats:
Total number of objects       : 23957
Number of monitored nodes :   1218
Number of Collectable DCIs  :11419

Lots of alarms are being created about node down and up. I don't want to change poll count for status change setting on server.

my server configuration about polling :

"Name","Value","Restart"
"NumberOfBusinessServicePollers","10","Yes"
"NumberOfConditionPollers","250","Yes"
"NumberOfConfigurationPollers","250","Yes"
"NumberOfDatabaseWriters","200","Yes"
"NumberOfDataCollectors","1000","Yes"
"NumberOfDiscoveryPollers","250","Yes"
"NumberOfRoutingTablePollers","250","Yes"
"NumberOfStatusPollers","1000","Yes"
"NumberOfTopologyTablePollers","250","Yes"
"NumberOfUpgradeThreads","10","No"

I am using a pc that 10gb of ram and 8 core cpus.

What do you offer me about this settings, and about node status down up alarms ?

by the way, i am getting lots of snmp connection lost errors.
and i saw that, syncher thread is not responding.

Hi
in treshold section, you must select "script" from function, samples=1, activation event and deactivation event as your choice.

Press Script Button and write :

dcival=GetDCIValue($node, FindDCIByName($node, "Status")); (Status=your DCI name)
if (dcival>0 && decimal <1) return true;
return false;

if you applied template to all nodes, then you must delete template in my opinion.

I did a mistake as you said and could solve this error only by deleting template.

Quote from: edward.borst on August 26, 2014, 01:13:08 PM
Hi Alex,

There seems a problem in the table script with the column reportId.
As I can see from the source that is is an hybernate uuid binary column I think it should be binary(16) instead of binary(255).
I have tried it (recreated the table) and now it is showing up in the console.

Only thing is that I can't render the results. I'm getting the following error in the nxreporting log:
Exception in thread "Thread-9" java.lang.IllegalArgumentException: 32768 > 16384
        at java.util.Arrays.copyOfRange(Arrays.java:2549)
        at com.radensolutions.reporting.infrastructure.impl.TcpConnector$SessionWorker.sendFileData(TcpConnector.java:232)
        at com.radensolutions.reporting.infrastructure.impl.TcpConnector$SessionWorker.run(TcpConnector.java:174)
        at java.lang.Thread.run(Thread.java:722)

I am getting same error after changing binary(255) to binary(16). I did not change anything else.

I had a look to output directory, and saw that report has been created. I tried to open it with jaspersoft studio, it succeeded and I could export it as pdf with in jaspersoft studio. But I can't do this in netxms console.

I hope, it will be corrected in ashort time.

Hi community!

I am using netxms for 1 month.
I have lots of (nearly 1100) routers around my country in my network and I am not an ISP
So, I am having problems about my nodes' speeds and statusses.

Now, these are I can do with this project:
Send e-mail to my ISP about node that is down and at the same time, send sms to my managers about node's status. (thanks for scripting and server side command)
I can see if there is a saturation on any node and can decide easily if the node needs more bandwidth.

These are what I saw that must be in netxms or I could not do these with netxms   (in my opinion)

* In world map view, all nodes' names are visible and this is a disadvantage on networks like mine (nearly all map of my country is full of my routers' names and not all of them can be seen). If an option can be set as SHOW ONLY STATUS INDICATOR, it will be more understandable for people who see my world map. And  if user can right click on this status indicator in world map, and see node's popup menu it will be more usable

* I tried to split Alarms to classes in dashboard. There can be an option here: I can see Object Tools menu in Alarm Browser screen, but I can't see them in dashboard's alarm view section. It will be more useful if it can be add here (for operator's usability).

* As I told above, I can send e-mail to my ISP and send sms to my managers with one command with object tools->server side command option in node's right click popup menu. A little speciality may be added here: Because of I can't use object tools in dashboard's alarm view, I can not send alarm id to my script as a variable. If it can be done so, I could add ack to my script and users could know if this alarm is noticed to ISP. Or if an event object can be add for acknowledgements (ack, sticky ack,..).

* An other thing is : There may be a class as Alarm in NXSL. It can be used for e-mail or sms informations to network operators. For example : If node's bandwidth usage is about max., and if this happens 50 times in a day, an auto message can be send to network operators. Or if SNMP fails about 10 times a day, it can be noticed to network operators by e-mail or sms, too.

I have about 1000 servers on my network around country and I will try to use nxagent on these servers. I have no idea about nxagent for now, But I am sure, it is very usefull too.

Thanks to everybody in this community. This is a really very good project.

Hi.
Is it possible to add object tool menu to dashboard's alarm viewer section's popup menu? It is possible to see object tools in alarm browser window but I created some dashboards and classified alarms for users. So, i am not showing all alarms to all users and I want them to run object tool commands in their dashboards.

Thank you.

thank you.