• General

  • Continuous "Current delay for tasks scheduled on thread pool SERVER..." messages

Hi,

Ive just changed my instance type on AWS and suddenly started having the following problems as you can see from the logs.
There are repeated Current delay for tasks scheduled on thread pool SERVER..." messages

the create_session.js request is timing out and the delay for the threadpool then continues to grow larger and larger after each thread in the threadpool appears to become blocked.

I'm thinking it could be something to do with the Java version as this seems to me to be the only thing that has changed and my local version on Mac OSX is still working fine with the same configuration.

Please can you check if my lightstreamer version and my java version are available.

Please let me know how I can resolve the issue.

My zipped ogfile at DEBUG level is attached.

Ive also see that a new version of LS is available - is there any guide on how to upgrade - I noticed that the adapter.xml file format has changed. Please can you let me know how to upgrade - my current adapters.xml file is attached. I'm using the adapter class:

<adapter_class>com.lightstreamer.adapters.remote.data.RobustNetworkedDataProvider</adapter_class>

How can I create an equivalent configuration for the latest LS release.

Thank you very much.
The log shows that all session establishment attempts are blocked and this is why the whole SERVER pool is exhausted.
In these cases the log suggests taking a thread dump, but it is highly likely that all the invocations to notifyUser are getting no answer.
What cannot be determined is whether there is a problem at network level on the communication between the Proxy Adapter and the Remote Adapter (the DORIANEX Metadata Adapter in this case) or there is a block inside the remote NotifyUser implementation.
But, since you are in an upgrading phase, the first suspect is that you have upgraded the Remote SDK without upgrading the Server and Proxy Adapter. In fact, the new versions of the Remote Adapter SDKs which ship with Lightstreamer 6 are not backward compatible. May you please check this first?

I confirm that porting your configuration to LS 6 is not straightforward.
Our general suggestion (as introduced in the readme file) is to start from an installation of LS 6 and manually re-apply all the customizations (usually on configuration files and Adapter deployment, but also the start script and other files in the conf directory may be involved).
We acknowledge that this may not be easy and are thinking of making some tool available.

The adapters.xml files have also undergone a few changes and this is particularly evident for the configuration of the Proxy Adapters. So, again, you should start from the templates in
DOCS-SDKs\adapter_remoting_infrastructure\doc\adapter_robust_conf_template
and reconfigure your proxies.

On the other hand, in terms of changes from the current deployment to the new one, we can resume them in the following way:
  1. In adapters.xml, you should replace the class names of the proxy adapters, i.e.
    com.lightstreamer.adapters.remote.metadata.RobustNetworkedMetadataProvider
    and
    com.lightstreamer.adapters.remote.data.RobustNetworkedDataProvider
    with the string "ROBUST_PROXY_FOR_REMOTE_ADAPTER" (same for both cases).
  2. You can then remove the old "ls-proxy-adapters.jar" without replacing it, because the proxy adapters are now part of the Server binaries.
  3. The second step for adapters.xml is adding the
    <classloader>log-enabled</classloader>
    line for all proxies, which you have already done.
  4. Then it is also mandatory an upgrade of the SDK for the remote adapter (which you perhaps have already done, as said above).
Thank you very much for the help.

my package.json file had * for the lightstreamer-adapter version. So when re-installing the server the adapter was upgrading thus making it incompatible with the server version of LS.

running npm install lightstreamer-adapter@1.0.2 resolved the issue.

We wold very much appreciate some tool to help the porting process, for now we will continue with our current setup which is otherwise working fine

Thanks again.