sashapolo commented on code in PR #5710:
URL: https://github.com/apache/ignite-3/pull/5710#discussion_r2063906876
##########
docs/_docs/administrators-guide/config/node-config.adoc:
##########
@@ -481,23 +506,6 @@ _Checkpointing_ is the process of copying dirty pages from
RAM to partition file
|aipersist.replacementMode|CLOCK|Sets the page replacement algorithm.| Yes |
Yes | CLOCK, RANDOM_LRU, SEGMENTED_LRU
|aimem.initSizeBytes|268435456| Initial memory region size in bytes, when the
used memory size exceeds this value, new chunks of memory will be allocated.|
Yes | Yes | Min 256Mb, max defined by the addressable memory limit of the OS
|aimem.maxSizeBytes|268435456| Maximum memory region size in bytes.| Yes | Yes
| Min 256Mb, max defined by the addressable memory limit of the OS
-|aimem.eviction.mode|DISABLED| Eviction mode.| Yes | No | DISABLED,
HISTORY_ONLY, RANDOM
-|aimem.eviction.threshold|90% a| Threshold for eviction initiation. A number
with a dimension identifier:
-
-* % - percentage of aimem.maxSize
-* k - Kb
-* m - Mb
-* g - Gb
-
-For instance, "90%" means that the page memory starts eviction only after 90%
of the data region is occupied.| Yes | No a| * 0-100%
-
-* 0-9223372036854775807k/m/g
-|aimem.eviction.lwmUpdateInterval|60000| Frequency of the low watermark update
in milliseconds.| Yes | No | 1 - inf
-|aimem.eviction.interval|60000| Interval between the data eviction
iterations.| Yes | No | 1 - inf
-|aimem.eviction.lwmThreshold|1000| If the low watermark is less than
evictionLwmThreshold from the current timestamp, the row eviction is
triggered.| Yes | No | 0 - inf
-|aimem.eviction.batchSize|60000| Eviction batch size in rows.| Yes | No | 1 -
inf
-|rocksdb.sizeBytes|268435456| Size of the rocksdb offheap cache.| Yes | Yes |
Min 0, max defined by the addressable memory limit of the OS
Review Comment:
Why did you remove this section?
##########
docs/_docs/administrators-guide/config/node-config.adoc:
##########
@@ -111,8 +111,47 @@ See the link:developers-guide/clients/overview[Clients]
section for information
|location|deployment|Relative path to folder in the working directory. All
deployment units content will be stored there.| Yes | No | A valid path
|======
+=== Failure Handler Configuration
+
+[source,json]
+----
+{
+ "ignite" : {
+ "failureHandler" : {
+ "dumpThreadsOnFailure" : true,
+ "dumpThreadsThrottlingTimeoutMillis" : 10000,
+ "handler" : {
+ "ignoredFailureTypes" : [
+ "systemWorkerBlocked", "systemCriticalOperationTimeout"
+ ],
+ "type" : "noop"
+ },
+ "oomBufferSizeBytes" : 16384
+ }
+ }
+}
+----
+
+[cols="1,1,3,1,1,2",opts="header", stripes=none]
+|======
+|Property|Default|Description|Changeable|Requires Restart|Acceptable Values
+
+|dumpThreadsOnFailure|true|The number of items that can be expired at once.|
Yes | Yes | true, false
+|dumpThreadsThrottlingTimeoutMillis|10000|Throttling timeout for thread dump
generation during failure handling, in milliseconds.| Yes | Yes | 1 - inf
+
+|handler.ignoredFailureTypes|[systemWorkerBlocked,
systemCriticalOperationTimeout]|Types of failures that will be ignored.| Yes |
Yes | 1 - inf
+|handler.type|noop|Failure handler configuration type.| Yes | Yes | noop
Review Comment:
Apart from the "noop" type, there are also `"stop"` and `"stopOrHalt"`
--
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]
