Process dies if profiling is turned on - tensorflow

I want to profile Tensorflow model on CloudML. When I use tf.RunOptions(trace_level=tf.RunOptions.FULL_TRACE), my process dies with non zero exit code without details of what happened.
I tried adding and removing the code which turns on this option, and there's 100% correlation between this option and the death of the process.
The error message is 'The replica master 0 exited with a non-zero status of 250. Termination reason: Error. To find out more about why your job exited please check the logs'
How can I diagnose and fix this problem?

It was fixed by using tensorflow 1.1.0 instead of 1.0.0. Though, profiling information wasn't shown.

For your question, basically the exit status means your code got a SIGABRT during run.
Update:
There is a issue of loading libcupti. Cloud ML Engine has found a bug related to it. Fix is in progress. The problem will be resolved in future release.

Related

How to solve the "Intervention Slow network is detected." error in React Native

I have been getting the following error:
"[Intervention] Slow network is detected. See https://www.chromestatus.com/feature/5636954674692096 for more details. Fallback font will be used while loading: http://localhost:19006/fonts/MaterialCommunityIcons.ttf"
While researching this error I have read a lot of people recommending to disable the error or just straight up ignore it, but since I have been getting this error the refresh time has increased substantially.
I had multiple instances running and hadn't restarted my PC in a while. After a restart it was fixed and back to loading almost instantly...

ResourceNotFoundException with full deploy to prod

I have a fully developed set of functions which work fine in the "dev" stage and it's now time for me deploy to production. Unfortunately every time I try to deploy it goes for a long time but after printing "Checking Stack update progress" it fails with a 404 error:
An error occurred: SentinelLambdaFunction - Function not found: arn:aws:lambda:us-east-1:837955377040:function:xyz-services-prod-sentinel (Service: AWSLambda; Status Code: 404; Error Code: ResourceNotFoundException; Request ID: 38f86b7a-99cd-11e8-af06-fffd92e40dc5).
This error is non-sensical to me as this function does exist and executing precisely the same full deployment to "dev" results in no error. Note that in both environment/stage, we are deploying 10 functions with a fully deployment.
I tried removing the function which was being complained about first, with the hope that I could re-include it on a second deployment but then it simply complained about a different function not existing.
I also thought maybe the "--force" parameter might push this deployment into place but it has had no impact on the error I get.
The cycle time for each attempt is very long so I'd be very grateful if anyone could help to point me in the right direction on this.
Below is a screenshot of the output when run in "verbose" mode:
In attempt to get around the error I thought maybe I'd have a better chance if I went into CloudFormation and explicitly deleted the template for prod. I attempted to do this from the GUI and got the following:
This actually has further convinced me that this removal is important but I'm not sure what to do next.
For me, the solution was:
serverless remove
and then try deploying again.
So the solution to this problem was to ensure that all previous traces of the CloudFront stack was removed. In my case I had manually taken out a few functions from Lambda and the 401 errors I was getting were likely occuring in the removal attempts rather than my assumption that it was related to adding these functions.
Bear in mind you may find yourself -- like I did -- where the first attempt to delete fails. In this case try again and make sure to check off any checkboxes exposed by UI that indicate what had caused the issues the prior attempt.
Once I'd done that I was able to deploy as per normal from the serverless framework.

ExecuteCaptureLog fails giving memory not available

I am processing SSAS Cube programmatically in .net script - server.ExecuteCaptureLog(true, true). It used to work perfectly. Suddenly now a days it started giving error saying
The operation has been cancelled because there is not enough memory available for the application. If using a 32-bit version of the product, consider upgrading to the 64-bit version or increasing the amount of memory available on the machine.
When we re-run it gives same error with couple of other errors.
Cursor not found
and
An error occurred while processing the partition 'partition1_0d0bbd49-7f8c-45c8-91ae-4y66a18b8cc22' in table 'partition1_0d0bbd49-7f8c-45c8-91ae-4y66a18b8cc22'".
Please help. Thanks in advance.
Ranjith

dispatching started for transformation

When I preview rows in Text file Input control of Pentaho, no rows appear and 'Show log' option displays this message
"Dispatching started for transformation".
What does it mean? How to overcome this issue?
It seems that either your transformation is invalid (you're missing one essential checkbox or another) or your PDI installation isn't working properly.
Which JAVA version are you using? And which PDI version? Try it on a fresh install and if it still doesn't work, go over your text file input step and validate that it's correctly configured.
Also, try removing all other steps, it could be that one of the subsequent steps is the one causing problems and stopping PDI from starting the transformation execution.
Well... maybe it's quite late, but I'm currently struggling with this issue in the Pentaho Community Version 8.
What I found, and solved some of my issues is that this message can be a potential warning for a Deadlock process. You have to be sure that none of this situations are present in your code:
An external component like a table lock by the database blocks the transformation.
The "Block this step until steps finish" step might run into a deadlock when there are more rows to process than the number of Rows in Rowset.
Within transformations there are situations when streams get split and joined again, so that the transformation blocks by design.
You could see full examples in the Jira Pentaho documentation page:
https://pentaho-community.atlassian.net/wiki/spaces/EAI/pages/386807182/Transformation+Deadlocks
I hope that it will help you!

Compiler internal error. Process terminated with exit code 3

I am using Intellij 9.0.4 I checked out a project from SVN, set up tomcat (its running locally), and now I am trying to Make (or Compile) but it keeps giving me Error:Compiler internal error. Process terminated with exit code 3. I have searched the internet and couldn't get this type exit code 3.
Do you have any idea? Or which log file should I have to check to see details of the problem?
Thanks
SOLVED: I got it. Just increase the maximum heap size of the compiler(Setting->Compiler->Java Compiler)
SOLVED: I got it. Just increase the maximum heap size of the compiler(Setting->Compiler->Java Compiler)