NetXMS Support Forum

This worked!

Many thanks!

I figured out that agent in daemon mode reacts with pain on sudo command in external script.
E.g. if script contain something

#!/bin/bash
echo 300

Than it works, but if we add sudo

sudo -u someuser echo 300

It does not. Did anyone run into this?

Hi All,

I'm using external parameters in nxagentd.conf file, one of them:

ExternalParameter = CF2PRIVATEMEM:privatemem.sh

If I launch nxagentd -f -D 9  it executes and reads the parameter correctly

[12-Aug-2016 14:30:59.404] [DEBUG] [session:0] Received message CMD_GET_PARAMETER
[12-Aug-2016 14:30:59.404] [DEBUG] [session:0] Requesting parameter "CF2PRIVATEMEM"
[12-Aug-2016 14:30:59.404] [DEBUG] H_ExternalParameter called for "CF2PRIVATEMEM" "Eprivatemem.sh"
[12-Aug-2016 14:30:59.404] [DEBUG] RunExternal called for "CF2PRIVATEMEM" "Eprivatemem.sh"
[12-Aug-2016 14:30:59.404] [DEBUG] RunExternal: command line is "privatemem.sh"
[12-Aug-2016 14:30:59.405] [DEBUG] RunExternal (shell exec): worker thread created
[12-Aug-2016 14:30:59.502] [DEBUG] H_ExternalParameter/POpenWorker: worker thread pipe read result: 0x7f8e8f5a5d70
[12-Aug-2016 14:30:59.503] [DEBUG] RunExternal (shell exec): execution status 0
[12-Aug-2016 14:30:59.503] [DEBUG] [session:0] GetParameterValue(): result is 0 (SUCCESS)
[12-Aug-2016 14:30:59.503] [DEBUG] [session:0] Sending message CMD_REQUEST_COMPLETED (size 56)
[12-Aug-2016 14:30:59.503] [DEBUG] [session:0] Outgoing message dump:
  ** 001D0000000000380000001800000002
  ** 0000001C000000000000000000000000
  ** 00000015010000000000000C00340037
  ** 0037003300310032
  ** code=0x001D (CMD_REQUEST_COMPLETED) flags=0x0000 id=24 size=56 numFields=2
  ** [    28] INT32    0
  ** [    21] STRING   "477312"


However if I launch in daemon mode nxagentd -d -D 9  then I get the following

[12-Aug-2016 14:29:31.930] [DEBUG] [session:0] Received message CMD_GET_PARAMETER
[12-Aug-2016 14:29:31.930] [DEBUG] [session:0] Requesting parameter "CF2PRIVATEMEM"
[12-Aug-2016 14:29:31.930] [DEBUG] H_ExternalParameter called for "CF2PRIVATEMEM" "Eprivatemem.sh"
[12-Aug-2016 14:29:31.930] [DEBUG] RunExternal called for "CF2PRIVATEMEM" "Eprivatemem.sh"
[12-Aug-2016 14:29:31.930] [DEBUG] RunExternal: command line is "privatemem.sh"
[12-Aug-2016 14:29:31.930] [DEBUG] RunExternal (shell exec): worker thread created
[12-Aug-2016 14:29:31.936] [DEBUG] RunExternal (shell exec): execution status 0
[12-Aug-2016 14:29:31.936] [DEBUG] [session:0] GetParameterValue(): result is 0 (SUCCESS)
[12-Aug-2016 14:29:31.936] [DEBUG] [session:0] Sending message CMD_REQUEST_COMPLETED (size 48)
[12-Aug-2016 14:29:31.936] [DEBUG] [session:0] Outgoing message dump:
  ** 001D0000000000300000001900000002
  ** 0000001C000000000000000000000000
  ** 00000015010000000000000000000000
  ** code=0x001D (CMD_REQUEST_COMPLETED) flags=0x0000 id=25 size=48 numFields=2
  ** [    28] INT32    0
  ** [    21] STRING   ""

Which is wrong. Current version of server and agent is 2.0.5 and it does not work here.
When I had version 1.2.17 it worked correctly.

Hello Victor,

Yes, you were correct I have removed nxmc directory under "/usr/share/tomcat6/work/Catalina/localhost" which is default for tomcat installation on RHEL. And this worked.

BR,
Roman

Solved it .

I usually rename nxmc-v.v.v.war to nxmc.war, so users can use the same URL like http://<IP>:8080/nxmc. And it didn't work with 1.2.7 version, if I live nxmc-1.2.7.war name everything works fine.

Hello Victor,

I've got trouble with new WEB GUI in 1.2.7. It displays an empty strings for the object tree, at the same time windows x86 client works fine. I've enclosed the screenshot of this.

Environment:
[root@netxms webapps]# cat /etc/redhat-release
CentOS release 6.2 (Final)
[root@netxms webapps]# rpm -qa | grep tomcat
tomcat6-lib-6.0.24-45.el6.noarch
tomcat6-servlet-2.5-api-6.0.24-45.el6.noarch
tomcat6-el-2.1-api-6.0.24-45.el6.noarch
tomcat6-6.0.24-45.el6.noarch
apache-tomcat-apis-0.1-1.el6.noarch
tomcat6-jsp-2.1-api-6.0.24-45.el6.noarch
[root@netxms webapps]#

Hi Victor,

The thing is that participants of this group can see and use everything under "objects" and nothing under "graphs" by default. (I have per group rights assigned on entities under "objects" and under "graphs" as well.) This is confusing, why not to have everything available to it.

Best regards,
Roman Davydchenko

Hello Victor,

This workaround works, many thanks for that!
I'll try web interface today later.

Best regards,
Roman

I run in error when compiling netxms on centos 6.3.

make[4]: Leaving directory `/distr/netxms-1.2.3.1/src/snmp/libnxsnmp'
Making all in nxmibc
make[4]: Entering directory `/distr/netxms-1.2.3.1/src/snmp/nxmibc'
  CXX    array.o
  CXX    lex.parser.o
  CXX    mibparse.o
  CXX    nxmibc.o
  CXX    parser.tab.o
  CXXLD  nxmibc
../../libnetxms/.libs/libnetxms.so: undefined reference to `tre_regwexec'
../../libnetxms/.libs/libnetxms.so: undefined reference to `tre_regwcomp'
collect2: ld returned 1 exit status
make[4]: *** [nxmibc] Error 1
make[4]: Leaving directory `/distr/netxms-1.2.3.1/src/snmp/nxmibc'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/distr/netxms-1.2.3.1/src/snmp'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/distr/netxms-1.2.3.1/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/distr/netxms-1.2.3.1'
make: *** [all] Error 2

Configure arguments "--with-server --with-pgsql --with-agent --with-snmp".

Yes, the predefined graphs tab (the one near the objects tab). He does not get any error, I see only "root" and nothing under it. If I add "admins" group to these graphs access rights, user will see the graphs. But the thing is I thought that "admins" group is like a predefined admin account that can view and alter anything. And in this case I have to define it for all objects, graphs, etc.

Best regards,
Roman

Hello Victor,

I ran in problem adding the second administrator. E.g. I create user and put him into admins group. And this user cannot view any graphs.
I using win32 console and netxms 1.2.2.

Best regards,
Roman

One thing I also would like to have is when you change alarm status to "Resolved" the alarm highlights should be removed from devices.

Hello Victor,

I have added SNMP string polling and it has the same effect as ODBC polling. Current polled value is good and stored are ruined.

Best regards,
Roman

Hello Victor,

Changing type to integer effectively solved the problem. All previously stored values are now displaying correctly. The one thing I don't understand is why did this happen?

Hello Victor,

I've managed to build and launch ODBC subagent tried it and got some issues. Current query returns right result, but those results stored previously die or something wrong happens to them. I have included the printscreens of management console showing this and screen of collection configuration.

Agent is configured with the following lines:
*ODBC
Query = SQL.MySQLQuery:mybase:SELECT COUNT(*) FROM USERS WHERE USERACTIVE = 1 AND USERDELETED = 0:60

Normally query should return 171 or something like this.