View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0004547 | SymmetricDS | Bug | public | 2020-09-14 19:59 | 2020-09-23 20:34 |
| Reporter | elong | Assigned To | elong | ||
| Priority | normal | ||||
| Status | closed | Resolution | fixed | ||
| Product Version | 3.12.2 | ||||
| Target Version | 3.12.4 | Fixed in Version | 3.12.4 | ||
| Summary | 0004547: Data Extractor gets "java.lang.IllegalStateException: There is no content to read" from staging | ||||
| Description | After optimizing concurrency in 0004472, a race condition was introduced where staging manager can remove a staged resource before it is push or pulled. When purging resources based on batches (stream.to.file.purge.on.ttl.enabled=false), the staging manager gets a list of batches from the database at the beginning of its run. As it looks through staged resources, if the resource does not have a corresponding batch from the list, it will remove the resource. The problem is that resources can be created while the job is in the middle of running, and those batches aren't on the list. To fix, we'll make a note of the start time of the job, and resources must have a last modified time before the job start time to be eligible for purging. 2020-09-11 17:00:00,007 INFO [corp-000] [StagingManager] [corp-000-job-18] Cleaning staging... 2020-09-11 17:00:00,050 INFO [corp-000] [StagingManager] [corp-000-job-18] Purged 224 staging files, freed 187 KB of disk space. 2020-09-11 17:00:00,050 INFO [corp-000] [StagingManager] [corp-000-job-18] Finished cleaning staging in 0 seconds. 2020-09-11 17:02:59,541 ERROR [corp-000] [DataExtractorService] [corp-000-push-default-5] Failed to extract batch 003-333923 StackTraceKey.init [IllegalStateException:427582167] java.lang.IllegalStateException: There is no content to read. Memory buffer was empty and /export/home/postgres/symmetric-ds/tmp/corp-000/outgoing/common/923/0000333923.create was not found. at org.jumpmind.symmetric.io.stage.StagedResource.getReader(StagedResource.java:246) at org.jumpmind.symmetric.service.impl.DataExtractorService.transferFromStaging(DataExtractorService.java:1481) at org.jumpmind.symmetric.service.impl.DataExtractorService.sendOutgoingBatch(DataExtractorService.java:1440) at org.jumpmind.symmetric.service.impl.DataExtractorService.extract(DataExtractorService.java:789) at org.jumpmind.symmetric.service.impl.DataExtractorService.extract(DataExtractorService.java:577) at org.jumpmind.symmetric.service.impl.PushService.pushToNode(PushService.java:206) at org.jumpmind.symmetric.service.impl.PushService.execute(PushService.java:167) at org.jumpmind.symmetric.service.impl.NodeCommunicationService$1.run(NodeCommunicationService.java:519) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) | ||||
| Steps To Reproduce | Add break point in debugger in BatchStagingManager to help with timing Run Stage Management job Change data and observe a new batch that is push/pulled Allow Stage Management job to continue, and it will immediately delete the new batch's stage file | ||||
| Tags | staging | ||||
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2020-09-14 19:59 | elong | New Issue | |
| 2020-09-14 19:59 | elong | Status | new => assigned |
| 2020-09-14 19:59 | elong | Assigned To | => elong |
| 2020-09-14 19:59 | elong | Tag Attached: staging | |
| 2020-09-14 19:59 | elong | Issue generated from: 0004546 | |
| 2020-09-14 19:59 | elong | Relationship added | related to 0004546 |
| 2020-09-14 19:59 | elong | Relationship added | related to 0004472 |
| 2020-09-14 20:01 | elong | Status | assigned => resolved |
| 2020-09-14 20:01 | elong | Resolution | open => fixed |
| 2020-09-14 20:01 | elong | Fixed in Version | => 3.12.4 |
| 2020-09-23 20:34 | admin | Status | resolved => closed |