Windows server 2016 license servers.

From server01 (TIMESTAMP 5/18/2026):

 3:20:07 (lmgrd) Attempting connection to primary redundant server server02

 3:20:29 (lmgrd) Disconnected from primary redundant server server02

 3:20:53 (lmgrd) Attempting connection to primary redundant server server02

 3:20:53 (lmgrd) Disconnected from primary redundant server server02

 3:21:15 (lmgrd) Attempting connection to primary redundant server server02

 3:21:19 (lmgrd) Established connection to primary redundant server server02

 3:21:19 (lmgrd) Disconnected from secondary redundant server server03

 3:21:19 (lmgrd) Redundant server quorum established. Master is server02

 3:21:19 (lmgrd) Starting vendor daemons ...

 3:21:19 (lmgrd) Starting vendor daemon at port 10306

 3:21:19 (lmgrd) Using vendor daemon port 10306 specified in license file

 3:21:19 (lmgrd) Started MLM (pid 4532)

server03: (TIMESTAMP 5/18/2026)

 3:21:19 (lmgrd) Disconnected from tertiary redundant server server01

 3:21:40 (lmgrd) Attempting connection to tertiary redundant server server01

 3:21:41 (lmgrd) Established connection to tertiary redundant server server01

server02: (TIMESTAMP 5/18/2026)

3:20:08 (lmgrd) Warning: Maximum connections to the server has reached. Please disconnect some clients from the server

...

3:21:11 (lmgrd) Redundant server heartbeat to server01 timed out

3:21:11 ((lmgrd)) Loop info: MT:3 VD_HB:58 reset:3 clients:1 fd:1 main:0

3:21:11 (lmgrd) Disconnected from tertiary redundant server server01

3:21:19 (lmgrd) Established connection from tertiary redundant server server01

There are many references to https://docs.revenera.com/fnp/2025r1/LicAdmin_Guide/Content/helplibrary/Maximum_Client_Connections_to_License_Server.htm to increase LM_SERVER_HIGHEST_FD but it does not seem relevant here clients:1 fd:1 suggests one client and 1 file descriptor? When server02 is looping there was a maximum of 300 ESTABLISHED TCP connections.

What is interesting is that "Maximum connections" message appears before heartbeat and loop info message on server02? So is "Maximum connections" a catch all error? If heartbeat connection is lost to redundant license servers and a license query is received from a client server02 has to queue as it cannot process when heartbeat is lost?, so in this case could a single query cause "Maximum connections" message to be logged as lmgrd knows due to missing heartbeats its not able to process requests?

Is "Maximum connections" message a catch all error? LM_SERVER_HIGHEST_FD link only seems relevant to genuinely busy license servers and doesn't seem to apply to us.

Up until failure occurred the vendor daemon was running on server02 (it was the master?), we don't understand server01 logs in same second server02 is identified as master it tried to also start the vendor daemon?

From the time the first "maximum connections" message is logged, to the "heartbeat timeout" being logged there is nominally a 60+ second time difference. So the question is setting HEARTBEAT_INTERVAL more relevant where network or server latency spikes might occur? (https://docs.revenera.com/fnp/2022r3/LicAdmin_Guide/Content/helplibrary/SERVER_Lines.htm).

Anyone have any ideas please, this is a long outstanding issue for us.

Best Regards - Colin

The attached FNP log file contains sections where multiple lines at the following time are not arranged in chronological order:

February 6, 2026, around 15:13: Lines 500703 to 500737

15:13:00 (SOFTWARE_PROD) UNSUPPORTED: "BIM" (PORT_AT_HOST_PLUS ) S21344@K000288178 (No such feature exists. (-5,346))

15:13:00 (SOFTWARE_PROD) UNSUPPORTED: "SHEETMETAL" (PORT_AT_HOST_PLUS ) S21344@K000288178 (No such feature exists. (-5,346))

--------From here-----------

15:12:42 (SOFTWARE_PROD) IN: "SOFTNW_PRO" S21344@K000288178

15:12:42 (SOFTWARE_PROD) IN: "SOFTNW_PRO" osk-t-00819w@K000291947

~~~

15:12:59 (SOFTWARE_PROD) OUT: "SOFTNW_PRO" W31459@K000270722

15:12:59 (SOFTWARE_PROD) UNSUPPORTED: "BIM" (PORT_AT_HOST_PLUS ) W31459@K000270722 (No such feature exists. (-5,346))

15:12:59 (SOFTWARE_PROD) UNSUPPORTED: "SHEETMETAL" (PORT_AT_HOST_PLUS ) W31459@K000270722 (No such feature exists. (-5,346))

---------Up to this point-----------

15:13:00 (SOFTWARE_PROD) UNSUPPORTED: "BIM" (PORT_AT_HOST_PLUS ) A22022@K000250591 (No such feature exists. (-5,346))

15:13:00 (SOFTWARE_PROD) UNSUPPORTED: "SHEETMETAL" (PORT_AT_HOST_PLUS ) A22022@K000250591 (No such feature exists. (-5,346))

If such logs are recorded, are there any explanation of this system behavior?

Hello FlexNet Support Team,

​

When using the lc_feat_list() API and the environment variable LM_LICENSE_FILE contains multiple license servers, if the valid server is not the first in the list, calling with the flag LM_FILTERLIST_ALL_FILES returns nullptr.

​

However, when I use LM_FLIST_ALL_FILES, the same call works correctly and lists features.

​

Could you please clarify the behavioral difference between LM_FILTERLIST_ALL_FILES and LM_FLIST_ALL_FILES?

• Does LM_FILTERLIST_ALL_FILES only query the first server in LM_LICENSE_FILE?
• Is this a known limitation or expected behavior?
• What is the recommended flag to ensure all license servers are checked?

​

Example:

LM_LICENSE_FILE=27000@serverA;27000@serverB;27000@serverC

​

Only serverB is valid for our vendor daemon:

• ​With LM_FILTERLIST_ALL_FILES: returns nullptr
• With LM_FLIST_ALL_FILES: works fine

​

Environment: FlexNet Publisher 11.19.0, Linux x86_64

We are using Flexnet publisher 11.19.6 for Java SDK. 

To access the trusted storage from a Java based application, we have to put the mandatory dll files for native access to the trusted storage into the working directory.

Loading the dll files is NOT the problem. 

They are properly found and loaded. 

But when checking out a feature, then we get the following result:

Error List Returned (-513,7) ; Native Method Call Failed (-530,3103) (tsGetFeatureLines: flxInitLoad failed.: 1); Error List Returned (-513,4050)

We already tried to change the current working directory to the folder containing the dlls. 

Then the checkout works properly.

However this is not suitable for our Eclipse RCP based Java Application.

Is there a way to tell the Flexnet SDK for Java applications, that the dlls are in a certain location?

Or get it somehow else to work?