iotdb icon indicating copy to clipboard operation
iotdb copied to clipboard

Published 20 hours ago • verified publisher icon apache

[IOTDB-1136] Improved reliability in flush error

Open wangchao316 opened this issue 4 years ago • 1 comments

The flushing fails due to OOM or insufficient disk space etc.... As a result, the cluster is in read-only mode. We need a mechanism to restore read-only mode. If the fault is caused by a common error

  1. If the flush fails, a retry mechanism can be added. By default, the retry mechanism is used for three times at an interval of 1s.
  2. After the read-only mode is enabled, the system can recover the read-only mode and pull the data every 10 minutes.

wangchao316 avatar Feb 20 '21 07:02 wangchao316

@qiaojialin @jixuan1989 hi, could you please review this? Thanks...

wangchao316 avatar Feb 22 '21 06:02 wangchao316