Hi Vannroath, It looks like you’re almost there.
NGINX and UWSGI: the taurus-api_00.log file contains the necessary clue: “Listen queue size is greater than the system max net.core.somaxconn (128)”. The nginx and uwsgi processes were bumping the TCP listen queue sizes to 1024. This was an unnecessary carryover from the "HTM for IT” showcase app. The REST API in Taurus is primarily for internal use (taurus.metric_collectors uses it), so we don’t need that big a queue size. The reason it failed on your system is that your system has the default net.core.somaxconn value of 128. I just merged https://github.com/numenta/numenta-apps/pull/695 into numenta-apps master that should address this issue by lowering the listen queue size configurations to 128. The other failure in your Taurus installation is characterized by the dynamodb-local-test-tool_00 log message: “ Error processing the supplied command line arguments: Invalid directory for database creation”. When running locally, Taurus uses an emulation tool called dynamodb-local-test-tool that emulates AWS’s DynamoDB service. Taurus’s supervisord.conf contains this command line for starting the dynamodb emulation tool: java -Djava.library.path=./DynamoDBLocal_lib -jar DynamoDBLocal.jar -port 8300 -dbPath %(ENV_HOME)s/.dynamodb "Invalid directory for database creation” likely means one of two things: 1. The home directory ($HOME) doesn’t have a subdirectory named .dynamodb (dot followed by dynomodb) where the emulation tool will store the emulated dynamodb items; or 2. The HOME environment variable is not defined in the environment from which the Taurus services are started or is set to an invalid value. Cheers, Vitaly From: nupic <[email protected]<mailto:[email protected]>> on behalf of Ky Vannroath <[email protected]<mailto:[email protected]>> Reply-To: "NuPIC general mailing list." <[email protected]<mailto:[email protected]>> Date: Sunday, March 13, 2016 at 10:03 AM To: Celeste Baranski <[email protected]<mailto:[email protected]>> Subject: Re: numenta-apps: taurus installation issue Hi Vitaly, Any update on my error? On Tue, Mar 8, 2016 at 3:34 PM, Vannroath Ky <[email protected]<mailto:[email protected]>> wrote: Hi Vitaly, I made the changes and able to start taurus service with supervisord in command prompt without error prompt. However, when i go to web interface at http://localhost:9001 I see the taurus:taurus-api_00 status: fatal Description: Exited too quickly (process log may have details). I tried to start the service in the web interface, but nothing happened. Am I missing something? taurus:taurus-api_00 Log: thunder lock: disabled (you can enable it with --thunder-lock) Listen queue size is greater than the system max net.core.somaxconn (128). *** Starting uWSGI 2.0.4 (64bit) on [Tue Mar 8 15:05:42 2016] *** compiled with version: 4.8.4 on 24 February 2016 15:38:37 os: Linux-3.19.0-51-generic #57~14.04.1-Ubuntu SMP Fri Feb 19 14:36:55 UTC 2016 nodename: meharu-VirtualBox machine: x86_64 clock source: unix detected number of CPU cores: 2 current working directory: /home/meharu/numenta-apps/taurus detected binary path: /usr/bin/uwsgi !!! no internal routing support, rebuild with pcre support !!! uWSGI running as root, you can use --uid/--gid/--chroot options *** WARNING: you are running uWSGI as root !!! (use the --uid flag) *** your processes number limit is 23790 your memory page size is 4096 bytes detected max file descriptor number: 1024 lock engine: pthread robust mutexes thunder lock: disabled (you can enable it with --thunder-lock) Listen queue size is greater than the system max net.core.somaxconn (128). dynamodb-local-test-tool:dynamodb-local-test-tool_00 log: Error processing the supplied command line argumentsInvalid directory for database creation Screenshot from supervisor web interface: [cid:[email protected]] Best, Vannroath On Tue, Mar 8, 2016 at 4:24 AM, Vitaly Kruglikov <[email protected]<mailto:[email protected]>> wrote: Hi Vannroath, I suspect that the environment variables `ENV_TAURUS_RMQ_METRIC_DEST` and `ENV_TAURUS_RMQ_METRIC_PREFIX` are not defined in your environment or something along those lines. When you grep in taurus for ENV_TAURUS_RMQ_METRIC_DEST, it takes you to taurus/conf/supervisord.conf. This command is part of the section "program:req-metric-collector”. This is a service that we run on our own production deployment of Taurus (HTM for Stocks). We use this service to send statistics about Taurus’s RabbitMQ instance (hence RMQ in the env var names) to our instance of HTM for IT, which monitors anomalies in the performance of Taurus’s message queues (among other IT metrics for our own servers). This helps us detect problems in our production deployments. You can remove the entire section "program:req-metric-collector” from your copy of supervisord.conf, since this service is nor required for Taurus to function. I recall from one of your earlier messages that you had a separate copy of the configuration directory (conf-user?) that you created for running Taurus. So, be sure to make your changes there. Best, Vitaly From: "Ky Vannroath" To: "'NuPIC general mailing list.'" <[email protected]<mailto:[email protected]>> Subject: RE: numenta-apps: taurus installation issue Message-ID: <[email protected]<mailto:[email protected]>> Content-Type: text/plain; charset="iso-8859-1" Hi Vitaly, I successfully initialized the database after changing what you instructed. Then I tried to follow thru the readme and face another issue on part "Start Taurus services with supervisord" with the following error. Please advise. root@meharu-VirtualBox:/home/meharu/numenta-apps/taurus# supervisord -c conf-user/supervisord.conf Error: Format string 'python -m htmengine.monitors.rmq_metric_collector_agent --metric-addr=%(ENV_TAURUS_RMQ_METRIC_DEST)s --metric-prefix=%(ENV_TAURUS_RMQ_METRIC_PREFIX)s' for 'command' contains names which cannot be expanded For help, use /usr/local/bin/supervisord -h Best, Vannroath
