Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
-
Reviewed
Description
While validating some HA failure modes we found that HDFS clients can take a long time to recover or sometimes don't recover at all since we don't setup the socket timeout in the InputStream:
private void transfer () { ...
...
OutputStream unbufOut = NetUtils.getOutputStream(sock, writeTimeout);
InputStream unbufIn = NetUtils.getInputStream(sock);
...
}
The InputStream should have its own timeout in the same way as the OutputStream.
Attachments
Attachments
Issue Links
- is duplicated by
-
HDFS-14587 Support fail fast when client wait ACK by pipeline over threshold
- Resolved
- is related to
-
HDFS-13103 HDFS Client write acknowledgement timeout should not depend on read timeout
- Open