Runbook a.k.a FAQ

• I am not getting any data from Remote Clusters for mirrored and merged Streams? Check databus.log and grep for MirroredStream/MergedStream and any exceptions you may see. Also check if you are falling into[2]

• Cluster was down for 1 day/network link was down for a day. After restore data got purged but it's still MergedStream/MirroredStream threads are trying to PULL? if retention period < PULL period then data is deleted from local cluster but still resides in the fileList to be pulled causing the PULLER thread to block as distcp fails if it can't find even one file among the list. Scenario - if retention is 1 day and network link is down for 1 day then purger will remove the data from streams and local_streams. To overcome open the file from the right path(explained in Paths page) and delete the file entries which don't exist. Databus threads will pull it in the next run.

• Can PRIMARY/MIRROR be same cluster for a given stream? No, this should never be the case. There are validations in DATABUS to avoid this.

• Is it safe to Start/Stop scribe collectors within the same minute.? No, it has a side effect of producing two files within the same minute. This also confuses DATABUS while moving this file to TRASH PATH. It can result in the same file being replayed with every new minute file. Recommendation - Any stop/starts of scribe collectors should be have 1 min sleep lag.