NetXMS Support Forum

Hi,

We have a number of DCIs using last( # ) samples equal to a specific value thresholds. I noticed that every time the NetXMS service restarts these thresholds are deactivated and reactivated # polls later if they were active before NetXMS went down.

Probably this has to do with the server not knowing what the previous values were because it was down. However, it creates the false impression that the threshold was no longer reached at a certain point.

Seeing as threshold activation/deactivation status is kept in the database, would it be possible to retain the old status of a threshold if the current DCI value(s) do not change it, and last values are not available due to a server restart?

For example, let's assume a DCI with threshold last(2) < 10. In the following list of polled values, assume X is the point where the server restarts:
15, 26, 30, X, 40, 50, 60

Currentely (tested on v1.2.6), the threshold would be active before X, be deactivated when it polls 40 first time after the restart, and be reactivated at value 50.
With the change I suggested, it would remain active throughout the test.

Another example:
15, 26, 30, X, 0, 1, 2
In this case, both the current version and my proposal should deactivate the threshold at 0.

15, 26, 30, X, 40, 0, 1
Here, the current version would deactivate the threshold on 40, but my proposal would delay this until the next poll (0).

Kind regards,
bdefloo

Hi,

The parameters available in event processing are listed here:
http://wiki.netxms.org/wiki/UM:Event_Processing#Macros_for_Event_Processing

You can use up to 99 variables linked to a particular event with the macros %1, %2, ..., %99

I can't seem to find any documentation of parameters for actions, but the principle is the same as parameters for ExternalParameter which are described in a stickied thread in the forums:
https://www.netxms.org/forum/configuration/agent-config-(action-externalparam)/msg3136/#msg3136

Hope this helps,

Kind regards,
bdefloo

Hi,

NetXMS uses switch forwarding tables to determine on what switch interface a node sits. This FDB contains the list of MAC addresses a switch knows on a particular interface.

As your workstation is connected to the phone, and the phone to the switch, both addresses appear on the same switch port in NetXMS. But, as it cannot read SNMP values from the IP phone as if it were a switch, it thinks there's another switch or hub in between. Because of this reason, it says they are indirectly connected.

As for the Layer 2 map, I believe it's having the same problem, and won't draw two devices directly connected to a single switch port. (But I'm not sure about this part, haven't really used the L2 maps alot)

Kind regards,
bdefloo

Hi!

Is the syslog functionality fixed in this version?

Any fixes relating to memory overflows/corrupt memory that could cause crashes?

Kind regards,
bdefloo

Hi,

Are your tablet and the NetXMS server in the same network? If the tablet is connecting directly to internet it can't communicate with an internal IP address such as 192.168.x.x.

If you are running the NetXMS server on Windows, you may need to disable the firewall or add an exception for incoming TCP connections on port 4701.

Kind regards,
bdefloo

Hi,

We've been having crashes similar to these for quite some time (from v1.2.0 up to v1.2.5) on windows 2003 x86.
I'm curious if we're having the same problem. Are there multiple events from the NetXMS Agent in the event log, starting minutes up to hours before the actual crash, with the message:
Communication session broken: An existing connection was forcibly closed by the mote host.
or does the server act strangely before crashing (slow response, node statuses changing to unknown, failing to show graphs/history, ...)?

In our case, I've never been able to pick anything useful out of the server logfile or the crash dumps, possibly because the logs fill up very quickly and memory corruption can take place long before the server actually tries to use the corrupted memory and crashes.

How often does your server crash? The rate at which the problem occurs seems tied in with queue lengths and memory usage in general. When we had large database writer queue lengths (> 100,000) the server was crashing 2-3 times a day, but after improving database performance somewhat (more database writers etc) it's now down to every 2-3 weeks.

I'm also hoping
https://www.netxms.org/forum/general-support/out-of-memory-netxms-v1-2-5/
is related to this problem, as on linux there's the chance that valgrind might catch the culprit.

Kind regards,
bdefloo

Hi,

That appears to have fixed it:

NetXMS Database Manager Version 1.2.5

Checking database (excluding collected data):
* Checking zone objects...                                            [PASSED]
Unlinked node object 31895 ("x") can be linked to subnet 93 (x.x.x.x). Link? (Y/N) Y
Unlinked node object 31963 ("x") can be linked to subnet 93 (x.x.x.x). Link? (Y/N) Y
Unlinked node object 62336 ("x") can be linked to subnet 9262 (x.x.x.x). Link? (Y/N) Y
* Checking node objects...                                            [FIXED ]
Unlinked interface object 36721 ("x"). Delete it? (Y/N) Y
Unlinked interface object 36722 ("x"). Delete it? (Y/N) Y
* Checking interface objects...                                       [FIXED ]
* Checking network service objects...                                 [PASSED]
* Checking cluster objects...                                         [PASSED]
* Checking template to node mapping...                                [PASSED]
* Checking object properties...                                       [PASSED]
* Checking event processing policy...                                 [PASSED]
5 errors was found, 5 errors was corrected
All errors in database was fixed
Commit changes? (Y/N) Y
Committing changes...
Changes was successfully committed to database
Database check completed

(Note: I removed hostnames/IPs)

After starting the server again and removing those nodes again, they were gone from the database and a subsequent nxdbmgr check didn't find any issues.

Thanks!

Hi,

Have you been able to find out more about this problem yet?

I have found a workaround for the first problem, namely just cancelling the "Select image" dialog immediately and opening it again. It only does the refreshes the first time it's opened after changing an image.

Kind regards,
bdefloo

Hi,

I've been having a strange issue on my v1.2.5 Windows 2003 x86 setup (but I think it's since v1.2.4):
I have a DCI on all my server nodes which checks if the last 5 ICMP ping results are above a certain value.
When the NetXMS core service starts, and one of these nodes is down, it will be triggered and post an event after 5 polls, which then sends an e-mail.

However, I appear to be getting these events and the related e-mails even for nodes I already deleted weeks ago! The node is nowhere to be found in the console object browser, and the alarms I'd expect to see aren't there either, but when I look in the database I can still find the deleted node in the object_properties and nodes tables, the DCIs are still there in the items table, ...
The is_deleted field is also still set to 0 in the object_properties table.

Could something be going wrong with the node deletion process?

Thanks in advance for any help you can provide!

Kind regards,
bdefloo

I'm having the same issue, running v1.2.5 on Windows 2003 x86. Syslog messages are received in the syslog log and monitor, but no subsequent events are generated from the parser. (can't find them in the event log/the actions aren't being executed)

Using log parser:
<parser>
<file></file>
   <rules>
<rule>
<match>(.*)</match>
<event params="1">100043</event>
</rule>
   </rules>
</parser>

event config:
<?xml version="1.0" encoding="UTF-8"?>
<configuration>
<formatVersion>3</formatVersion>
<description></description>
<events>
<event id="100043">
<name>Operations Warning Issue (Log parser)</name>
<code>100043</code>
<severity>1</severity>
<flags>1</flags>
<message>Warning issue with %n: %1.</message>
<description></description>
</event>
</events>
<templates>
</templates>
<traps>
</traps>
</configuration>

With debugging level 9 on, I get the message
[29-Jan-2013 09:34:41.030] syslogd: parser successfully created from config
when saving the syslog configuration, but nothing appears concerning actual syslog messages. I expected to see an event like
Match event: source=\"%s\" id=%u level=%d text=\"%s\"
when the syslog message arrives in the parser.

Might be unicode-related? Seems like the only thing changed recently to the syslog code.

Hope you can find something!

Hi,

My two posts are two mostly unrelated problems. I probably should've posted them in a seperate topic.

I just recreated the problem in the first post as follows:
I have an image library of some 30 images. I add another image. When I then open the properties of a map, and click "..." on the Map background page, I get the "Select image" dialog. All the images are shown in it within a second or two. However, it reloads several times after that, the scrollbar each time jumping back to the top.
(Note: I only just now discovered that something has to have changed in the image library before this behaviour appears.)

Second problem:
Suppose I have 2 maps, A and B, each with a background image. I have both maps open in the console, and in the image library right click the background image for map A and select edit. Then, I select a different image file using the "..." button and click OK.
Now, I switch back to the tab with map A. It's now using the newly selected background image. Switch to tab B, it's no longer showing anything (or just some remnants from the last tab I was looking at)

I'm using PNG images, the ones I used for these test were 752x480, about 900k ~ 1MB in size.

Hope this helps you to track down the issues!

Kind regards,
Bastiaan

Found another possible bug; someone may need to confirm it.

If you change an image in the image library while a network map is open, and you're not changing that map's background image, the map needs to be closed and reopened before you can view it.

I'm having issues editing existing map backgrounds in the image library and uploading new images over them. The map I uploaded a new image for still views fine, ones that I didn't touch don't. This in v1.2.5 with the x64 Windows console.

Hi,

Thanks for fixing the bug in the Image library that caused it to refresh over and over whenever something changed, once for every image in the library. However, the same thing seems to be happening when opening the "Select Image" dialog, under for example Map Properties->Map background tab, causing an unneccessary delay in selecting an image.

Keep up the great work!

Kind regards,
bdefloo

Hi,

Just for future reference of anyone having the same problems as we did upgrading to v1.2.5 on MS SQL 2008. We upgraded from v1.2.4 on a 32bit Windows 2003 server. When upgrading the database I got the following error:
C:\NetXMS\bin>nxdbmgr.exe upgrade
NetXMS Database Manager Version 1.2.5

Upgrading database...
Upgrading from version 265 to 266
SQL query failed ([Microsoft][SQL Server Native Client 10.0][SQL Server]Cannot c
reate index. Object 'event_log' was created with the following SET options off:
'ANSI_NULLS'.):
CREATE INDEX idx_event_log_root_id ON event_log(root_event_id) WHERE root_event_
id > 0
Rolling back last stage due to upgrade errors...
Database upgrade failed


Most likely, this was because we've been upgrading the same NetXMS instance since v1.0.13 or older. The only solution we found was to rename the event_log table, and to recreate it with ANSI_NULLS on:
USE [netxms]
GO

/****** Object:  Table [netxms].[event_log]    Script Date: 01/11/2013 10:37:06 ******/
SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

SET ANSI_PADDING OFF
GO

CREATE TABLE [netxms].[event_log](
[event_id] [bigint] NOT NULL,
[event_code] [int] NOT NULL,
[event_timestamp] [int] NOT NULL,
[event_source] [int] NOT NULL,
[event_severity] [int] NOT NULL,
[event_message] [varchar](255) NULL,
[root_event_id] [bigint] NOT NULL,
[user_tag] [varchar](63) NULL,
PRIMARY KEY CLUSTERED
(
[event_id] ASC
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]
) ON [PRIMARY]

GO

SET ANSI_PADDING OFF
GO


Probably you could then move all the data back to the new table, but we didn't really have any need for it so we just started with an empty table.

Hope this helps anyone coming across the same issue.

Kind regards,
bdefloo

Sounds great, thanks!