NetXMS Support Forum

Hi,

I just fixed that. Fix will be included into next patch release.

Best regards,
Victor

Hi,

version 3.0 hides interfaces in object tree by default (this significantly improves UI performance on large number of nodes). You can switch back to old behavior by unchecking "hide node components" checkbox in object tree view menu.

Best regards,
Victor

Hi,

looks like a bug in UI. What version you are using? Is it web or desktop UI?

Best regards,
Victor

Hi,

is there some kind of deadlock logging in MariaDB? This can help if we can find out what transaction holds the lock. If it was not logged I think we have to wait for next occurrence and then if I understand correctly command SHOW ENGINE INNODB STATUS should show active transactions and their locks which can provide useful information.

Best regards,
Victor

Hi!

Please attach console log (it should be in your user's home folder - C:\Users\<username>\.nxmc\data\.metadata\.log).

Best regards,
Victor

That could be memory leak. Do you have graph of memory consumption by netxmsd process?

Best regards,
Victor

Hi,

could you please post full explains for same queries run on idata and on idata_sc_default?

Best regards,
Victor

Check this thread: https://www.netxms.org/forum/installation/netxms-3-0-2292-configure-error-cannot-find-mysql-h-on-centos7/

Hi,

looks like you have to install MariaDB-shared as well. Then --with-mysql should work. I also changed configure to provide option for setup like this when MariaDB library tries to mimic MySQL client library and installs headers into /usr/include/mysql: --with-mariadb-compat-headers. You can try this source package: https://netxms.org/download/releases/3.0/netxms-3.0.2307.tar.gz and use configure options --with-mariadb --with-mariadb-compat-headers.

Best regards,
Victor

Hi all,

we have just published NetXMS 3.0 patch release 3 (version number 3.0.2305). It affects only server part and contains the following fixes:

- Fixed bug in setting string values via SNMP SET
- Fixed bug in Oracle database driver
- Fixed bug in database upgrade procedure
- Fixed bug in nxnotify command line parsing

Best regards,
Victor

Должно быть %%

Hi,

now you should use state attribute to determine node state. For convenience server version 3.0 defines constants for each state:

NodeState::Unreachable
NodeState::NetworkPathProblem
NodeState::AgentUnreachable
NodeState::SNMPUnreachable
NodeState::CacheModeNotSupported

ClusterState::Unreachable
ClusterState::NetworkPathProblem
ClusterState::Down

SensorState::Unreachable
SensorState::NetworkPathProblem
SensorState::Provisioned
SensorState::Registered
SensorState::Active
SensorState::PendingConfigUpdate

So equivalent of 2.2.x

if ($node->runtimeFlags & 0x0004)


will be


if ($node->state & NodeState::Unreachable)


State bit 4 is not used and it's presence is a bit of a mystery. Probably something goes wrong during DB upgrade. You can left it as is - it will not affect anything, or clean it up in database while server is stopped using the following queries:

UPDATED object_properties SET state=0 WHERE state=4;
UPDATED object_properties SET state=1 WHERE state=5;

This will clean it up for most common flag combinations.

If you are using PostgreSQL you should be able to do this in one query using bitwise AND:

UPDATED object_properties SET state=state & ~4;

Best regards,
Victor

Hi,

there is no built-in function specifically for FDB retrieval but you can do SNMP walk and retrieve FDB that way. I see more problems with calling external REST API. Alternatively you can do everything from nxshell script - get FDB, do REST call, and rename interface.

Best regards,
Victor

Hi,

we don't have other reports for such behavior, even for quite large deployments. Please turn on debug logging on proxy and server to level 6 and send part of the logs for period when one such issue experienced.

Best regards,
Victor

Это проблема старого агента - он возвращал неполные версии пакетов и получались дубликаты. Обновление агентов должно решить проблему.