AppManage Initialization Slow

[Tracy Martin]

We see differences in how long it takes to deploy an application to our UAT and PROD enviroments. Our UAT and PROD environments are both VMs running on the same specifications, same operating system, and same version of TIBCO TRA (5.10.1_V4). I have not been able to find any differences.

AppManage output shows the following except the first time column which is a calculation of the time difference between the previous line and current line:

 

 

 

0:00:00

01:55:31

Initializing ...

 

 

 

 

 

0:00:01

01:55:32

244 [main] INFO com.tibco.security.impl.np.SecurityVendor - Initializing Entrust crypto provider in NON FIPS 140-2 mode; insert provider as normal

 

 

0:01:42

01:57:14

INFO : Connected to: tcp://AUAPGPDTIBAPP03.GENPT.NET:7222 ## Thu Apr 09 01:57:14 EDT 2020 ##

 

 

 

0:00:01

01:57:15

INFO : Connected to: tcp://AUAPGPDTIBAPP03.GENPT.NET:7222 ## Thu Apr 09 01:57:15 EDT 2020 ##

 

 

 

0:00:00

01:57:15

Finished initialization

 

 

 

 

 

 

In this example you can see it took 1:42 to for one portion of the initialization process. In UAT, it takes 5 seconds. This is not a huge problem for 1 instance, but we when are deploying 70 instances, it is a bigger issue. Our deployment stops an instance and starts a new instance for each copy of the application running in a round-robin fashion so that we do not have downtime. We typically have 2 or 4 instances for a given application. After all of the new instances have been started, we undeploy the old instance. I calculated a total of 3 hours difference between our production and UAT deployments within the "Initializing ..." and "Finished initialization" steps.

We are exploring using batches to deploy, but that will give us some downtime, which we would prefer not to do.

My question is: Does anyone know what could be causing this slowness during the initialization step of AppManage

[Emmanuel Marchiset 2]

Hi, This can be for different reasons... delays in connecting to the EMS server or the domain database possibly because of issues at network level, delays at OS level (you have to check the ulimit set-up with ulimit -a command.

Also you may face a know issue with Java under Linux :

https://ruleoftech.com/2016/avoiding-jvm-delays-caused-by-random-number-generation