ptlrs commented on code in PR #9954:
URL: https://github.com/apache/ozone/pull/9954#discussion_r3114335525
##########
hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/common/statemachine/DatanodeConfiguration.java:
##########
@@ -430,6 +441,39 @@ public class DatanodeConfiguration extends
ReconfigurableConfig {
)
private Duration diskCheckSlidingWindowTimeout =
DISK_CHECK_SLIDING_WINDOW_TIMEOUT_DEFAULT;
+ @Config(key = "hdds.datanode.disk.check.timeout.test.enabled",
+ defaultValue = "true",
+ type = ConfigType.BOOLEAN,
+ tags = { DATANODE },
+ description = "Enable or disable timeout-based volume failure tracking "
+ + "independently of the disk IO health checks."
+ )
+ private boolean isDiskCheckTimeoutTestEnabled =
+ DISK_CHECK_TIMEOUT_TEST_ENABLED_DEFAULT;
+
+ @Config(key = "hdds.datanode.disk.check.timeout.failures.tolerated",
+ defaultValue = "1",
+ type = ConfigType.INT,
+ tags = { DATANODE },
+ description = "The number of volume check timeouts allowed within the "
+ + "timeout sliding window before the volume is marked as failed."
+ )
+ private int diskCheckTimeoutFailureTolerance =
+ DISK_CHECK_TIMEOUT_FAILURES_TOLERATED_DEFAULT;
+
+ @Config(key = "hdds.datanode.disk.check.timeout.sliding.window.timeout",
+ defaultValue = "70m",
Review Comment:
@jojochuang As the disk checks run every 1 hour and in worst case take 10
minutes, the number is 70 minutes to accommodate two runs because we want to
fail if two consecutive runs fail.
However, I think it needs to be 80 minutes because the executor service
maintains a 60 minute gap between consecutive runs as it is of type
`scheduleWithFixedDelay`.
Maybe let's make it a round number to 90 minutes or 1.5 hours.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]