If you see an error the one below this in the AEM logs then it is likely you are experiencing a bug with Datastore GC cleaning up files that are in use. If you are using a FileDataStore then see this article for details on how to repair the corruption.
10.09.201510:38:04.220*ERROR* [pool-9-thread-3] org.apache.sling.commons.scheduler.impl.QuartzScheduler Exception during job execution of org.apache.jackrabbit.oak.plugins.index.AsyncIndexUpdate@7fe7e8fa: Error occurredwhileobtaining InputStreamforblobId [add1dd8fc5093b27b1fae1b753cb48b24ef3231f#101983]
No matter whatever you do , very difficult to get rid of this issue. If its lower environment or local instance create a fresh instance and reinstall the content and code. If its production you can follow the process to recover.
I am suggesting to recreate the instance for Lower and local environment - Because it will save lot of time.