[
https://issues.apache.org/jira/browse/KAFKA-8733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17281204#comment-17281204
]
Ajay Patel commented on KAFKA-8733:
-----------------------------------
Is there any update on this or on
[KIP-501|https://cwiki.apache.org/confluence/display/KAFKA/KIP-501+Avoid+out-of-sync+or+offline+partitions+when+follower+fetch+requests+not+processed+in+time]?
Unfortunately the {{replica.lag.time}} workaround isn't exactly viable for us
as we predominantly see this issue when certain zones of our clusters have
complete outages. Enabling unclean leader election certainly resolves the
issue, but leaving the configuration on by default poses a significant risk of
data loss that we'd like really like to avoid. Thanks!
> Offline partitions occur when leader's disk is slow in reads while responding
> to follower fetch requests.
> ---------------------------------------------------------------------------------------------------------
>
> Key: KAFKA-8733
> URL: https://issues.apache.org/jira/browse/KAFKA-8733
> Project: Kafka
> Issue Type: Bug
> Components: core
> Affects Versions: 1.1.2, 2.4.0
> Reporter: Satish Duggana
> Assignee: Satish Duggana
> Priority: Critical
> Attachments: weighted-io-time-2.png, wio-time.png
>
>
> We found offline partitions issue multiple times on some of the hosts in our
> clusters. After going through the broker logs and hosts’s disk stats, it
> looks like this issue occurs whenever the read/write operations take more
> time on that disk. In a particular case where read time is more than the
> replica.lag.time.max.ms, follower replicas will be out of sync as their
> earlier fetch requests are stuck while reading the local log and their fetch
> status is not yet updated as mentioned in the below code of `ReplicaManager`.
> If there is an issue in reading the data from the log for a duration more
> than replica.lag.time.max.ms then all the replicas will be out of sync and
> partition becomes offline if min.isr.replicas > 1 and unclean.leader.election
> is false.
>
> {code:java}
> def readFromLog(): Seq[(TopicPartition, LogReadResult)] = {
> val result = readFromLocalLog( // this call took more than
> `replica.lag.time.max.ms`
> replicaId = replicaId,
> fetchOnlyFromLeader = fetchOnlyFromLeader,
> readOnlyCommitted = fetchOnlyCommitted,
> fetchMaxBytes = fetchMaxBytes,
> hardMaxBytesLimit = hardMaxBytesLimit,
> readPartitionInfo = fetchInfos,
> quota = quota,
> isolationLevel = isolationLevel)
> if (isFromFollower) updateFollowerLogReadResults(replicaId, result). //
> fetch time gets updated here, but mayBeShrinkIsr should have been already
> called and the replica is removed from isr
> else result
> }
> val logReadResults = readFromLog()
> {code}
> Attached the graphs of disk weighted io time stats when this issue occurred.
> I will raise [KIP-501|https://s.apache.org/jhbpn] describing options on how
> to handle this scenario.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
