-
Notifications
You must be signed in to change notification settings - Fork 16
[$350] Duplicate tickets #234
Comments
Contest https://www.topcoder-dev.com/challenges/30107517 has been created for this ticket. |
Contest https://www.topcoder-dev.com/challenges/30107518 has been created for this ticket. |
@afrisalyp - Another one where duplicates were created: |
@jmgasper So far, I can reproduce the problem if I run multiple instances of the processor. Ideally, all of the instances must have the same KAFKA_GROUP_ID value. But I got some errors. So I ran multiple processors but with different KAFKA_GROUP_ID. This deployment makes the duplicate issue occurs. So I believe this is the main problem that causes this duplicate issue. |
@afrisalyp - I'm checking on that, but I'm not sure that's the issue. For instance, I closed out a ticket that got migrated over, and it's attempting to create a new challenge, instead of using the existing one. Can you check why that may be? Relevant logs:
|
@afrisalyp - It just seems like it's trying to create a new ticket for every action, like label assigned, user assigned, etc... |
For this particular ticket, I see multiple DynamoDB entries. I'm not sure why the initial ticket wasn't used for assigning:
|
@afrisalyp - Got confirmation that there is only one processor instance running. |
@jmgasper |
@afrisalyp - I don't think we've seen it in dev, no. I'll test to make sure. |
@afrisalyp - Is dev loading for you? |
@jmgasper |
@afrisalyp - Here's a sample from a test ticket I just logged.
|
Seems like we get |
@jmgasper |
This issue seems affect the prod only. The dev and my local deployment don't. |
@afrisalyp - Yeah, agreed. We didn't get this in dev in my testing. |
I sent logs targeting this test issue: #242 , and a full dump of the issues table via email. If you want to do extra logging anywhere, feel free to open a merge request with the additional logging and I'll get it deployed. |
@afrisalyp - Any ideas? Any additional logging we can add that would be helpful? |
@jmgasper |
@afrisalyp - Thanks. Deploying now. |
@afrisalyp - Where are the different Seems like when |
And maybe we have to add handling for an empty result? Seems like we don't have that currently. |
@jmgasper But I think we need to investigate the log related to |
I'm wondering how's the prod deployed. How's the connection between the topcoder-x app and Dynamodb instance? |
I put |
@afrisalyp - The environment is set up separately for AWS - this was discussed quite a bit with Topcoder when they made the change. That shouldn't affect the app randomly though, should it? If that was wrong, the app just wouldn't work at all. |
I'm not sure. But since I don't experience the duplication issue in my local env and also we don't see it happens in the dev env so far, I just think about the prod deployment. |
@afrisalyp - Yeah, the data was migrated, but I don't see anything weird in DynamoDB, do you? I'm also asking Topcoder about differences in how the DynamoDB tables are set up in dev vs. prod. I agree that it's really strange that this issue is only affecting dev. |
The changes to handle the null record seem to have improved things slightly. |
Contest https://www.topcoder.com/challenges/30108143 has been created for this ticket. |
This ticket was not processed for payment. If you would like to process it for payment, please reopen it, add the |
Contest https://www.topcoder.com/challenges/30108143 has been updated - it has been assigned to afrisalyp. |
Payment task has been updated: https://software.topcoder-dev.com/review/actions/ViewProjectDetails?pid=30108143 |
@afrisalyp - As you noted last week, we can sometimes get duplicate tickets, like what I saw here with the timeouts:
#233
We need to make sure that this can't happen. I'm open to suggestions.
The text was updated successfully, but these errors were encountered: