A Couple of Problems

[brucecui]

I've met two problems during my first tasting today:

1. It looks like the river can't get back working after delete and recreate it. I have to delete the entire _river index to make it work. Not sure if it's the current ES design or a defect of ES or jdbc plugin. If I have other rivers running, guess this will be a problem.
2. I still get " Can not write a field name, expecting a value" exception in the log. This happens when I use a result set with columns which could contain null values. The exception disappeared when I narrowed down the result with only some key columns which are not null. I see a fix pull regarding this issue, and I have downloaded the code as well. I built the code and uploaded the package to the cluster, but the error is still there. But since there is no line number mentioned in the log, I can't tell where caused this problem.

[jprante]

1. Which version of ES do you use? Can it be related to River name re-use not possible between node shutdowns elastic/elasticsearch#1921 ?
2. You are right, there are flaws with null values, I currently work on that issues and an overhauled version, and I will put on some more rigid testing. If you don't mind, it would be great if you could isolate the row data causing the exception and submit it here for building a test case.

[brucecui]

Thanks for the prompt response.

1. I checked the linked issue. Looks like it's not my case. I tries several times today but couldn't recur it. I guess it's because I didn't install the plugin on every node in the first place, but later I put it on all my three nodes since I found ES balanced the importing operation to other nodes and failed due to lacking of the plugin. Anyway if I meet this issue again and also can recur it I'll post here again.
2. It'll be very nice once we have the robustness on null values. Unfortunately the data I used were from the dev env of my current project, which contains some business sensitive data that I may not be allowed to publish here. But I'll try to locate the pain point and hopefully that will help.

[jprante]

JDBC river 2.0.0 released