[jira] [Assigned] (KAFKA-16899) Rename MembershipManagerImpl's rebalanceTimeoutMs for clarity
[ https://issues.apache.org/jira/browse/KAFKA-16899?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Linu Shibu reassigned KAFKA-16899: -- Assignee: Linu Shibu > Rename MembershipManagerImpl's rebalanceTimeoutMs for clarity > - > > Key: KAFKA-16899 > URL: https://issues.apache.org/jira/browse/KAFKA-16899 > Project: Kafka > Issue Type: Improvement > Components: clients, consumer >Affects Versions: 3.8.0 >Reporter: Kirk True >Assignee: Linu Shibu >Priority: Minor > Labels: kip-848-client-support > Fix For: 4.0.0 > > > The naming of {{{}MembershipManagerImpl{}}}'s {{rebalanceTimeoutMs}} is a > little confusing. > Ultimately the broker enforces the rebalance process' timeout, not the > client. It's therefore a little misleading to use {{rebalanceTimeoutMs}} as > the name of a variable that only handles the client's commit portion of the > process. It is used in {{MembershipManagerImpl}} as a means to limit the > client's efforts in the case where it is repeatedly trying to commit but > failing. > A suggested name change was {{{}commitTimeoutDuringReconciliation{}}}. > It is not clear to me that simply changing the name from > {{rebalanceTimeoutMs}} to {{commitTimeoutDuringReconciliation}} is > sufficient. When a caller instantiates a {{{}MembershipManagerImpl{}}}, it > is passing in a variable named {{rebalanceTimeoutMs}} that ultimately comes > from the {{ClientConfig}} object. I can imagine it being a little confusing > that a value for {{rebalanceTimeoutMs}} is passed in, but then really only > used as the value for {{commitTimeoutDuringReconciliation}}. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-16899) Rename MembershipManagerImpl's rebalanceTimeoutMs for clarity
[ https://issues.apache.org/jira/browse/KAFKA-16899?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17852591#comment-17852591 ] Linu Shibu commented on KAFKA-16899: Could I pick this up [~kirktrue] if you are not working on this? > Rename MembershipManagerImpl's rebalanceTimeoutMs for clarity > - > > Key: KAFKA-16899 > URL: https://issues.apache.org/jira/browse/KAFKA-16899 > Project: Kafka > Issue Type: Improvement > Components: clients, consumer >Affects Versions: 3.8.0 >Reporter: Kirk True >Priority: Minor > Labels: kip-848-client-support > Fix For: 4.0.0 > > > The naming of {{{}MembershipManagerImpl{}}}'s {{rebalanceTimeoutMs}} is a > little confusing. > Ultimately the broker enforces the rebalance process' timeout, not the > client. It's therefore a little misleading to use {{rebalanceTimeoutMs}} as > the name of a variable that only handles the client's commit portion of the > process. It is used in {{MembershipManagerImpl}} as a means to limit the > client's efforts in the case where it is repeatedly trying to commit but > failing. > A suggested name change was {{{}commitTimeoutDuringReconciliation{}}}. > It is not clear to me that simply changing the name from > {{rebalanceTimeoutMs}} to {{commitTimeoutDuringReconciliation}} is > sufficient. When a caller instantiates a {{{}MembershipManagerImpl{}}}, it > is passing in a variable named {{rebalanceTimeoutMs}} that ultimately comes > from the {{ClientConfig}} object. I can imagine it being a little confusing > that a value for {{rebalanceTimeoutMs}} is passed in, but then really only > used as the value for {{commitTimeoutDuringReconciliation}}. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Assigned] (KAFKA-16834) add the reason for the failure of PartitionRegistration#toRecord
[ https://issues.apache.org/jira/browse/KAFKA-16834?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Linu Shibu reassigned KAFKA-16834: -- Assignee: Linu Shibu > add the reason for the failure of PartitionRegistration#toRecord > > > Key: KAFKA-16834 > URL: https://issues.apache.org/jira/browse/KAFKA-16834 > Project: Kafka > Issue Type: Wish >Affects Versions: 3.7.0 >Reporter: Jianbin Chen >Assignee: Linu Shibu >Priority: Minor > > Transform it into the following output, which is easier for users to > understand and identify the cause of the problem. > {code:java} > options.handleLoss("the directory " + (directory == DirectoryId.UNASSIGNED ? > "unassigned" : "lost") > + " state of one or more replicas");{code} -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-16834) add the reason for the failure of PartitionRegistration#toRecord
[ https://issues.apache.org/jira/browse/KAFKA-16834?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17850163#comment-17850163 ] Linu Shibu commented on KAFKA-16834: [~jianbin] can I pick up and work on this? > add the reason for the failure of PartitionRegistration#toRecord > > > Key: KAFKA-16834 > URL: https://issues.apache.org/jira/browse/KAFKA-16834 > Project: Kafka > Issue Type: Wish >Affects Versions: 3.7.0 >Reporter: Jianbin Chen >Priority: Minor > > Transform it into the following output, which is easier for users to > understand and identify the cause of the problem. > {code:java} > options.handleLoss("the directory " + (directory == DirectoryId.UNASSIGNED ? > "unassigned" : "lost") > + " state of one or more replicas");{code} -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-16356) Remove class-name dispatch in RemoteLogMetadataSerde
[ https://issues.apache.org/jira/browse/KAFKA-16356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17830781#comment-17830781 ] Linu Shibu commented on KAFKA-16356: Yes [~yondy] , I will update this soon, thanks! > Remove class-name dispatch in RemoteLogMetadataSerde > > > Key: KAFKA-16356 > URL: https://issues.apache.org/jira/browse/KAFKA-16356 > Project: Kafka > Issue Type: Task > Components: Tiered-Storage >Affects Versions: 3.7.0 >Reporter: Greg Harris >Assignee: Linu Shibu >Priority: Trivial > Labels: newbie > > The RemoteLogMetadataSerde#serialize receives a RemoteLogMetadata object, and > has to dispatch to one of four serializers depending on it's type. This is > done by taking the class name of the RemoteLogMetadata and looking it up in > maps to find the corresponding serializer for that class. > This later requires an unchecked cast, because the RemoteLogMetadataTransform > is generic. This is all type-unsafe, and can be replaced with type-safe > if-elseif-else statements that may also be faster than the double-indirect > map lookups. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-16356) Remove class-name dispatch in RemoteLogMetadataSerde
[ https://issues.apache.org/jira/browse/KAFKA-16356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17826075#comment-17826075 ] Linu Shibu commented on KAFKA-16356: Thanks [~gharris1727]! I have assigned it to myself and will be working on it [~yondy] . > Remove class-name dispatch in RemoteLogMetadataSerde > > > Key: KAFKA-16356 > URL: https://issues.apache.org/jira/browse/KAFKA-16356 > Project: Kafka > Issue Type: Task > Components: Tiered-Storage >Affects Versions: 3.7.0 >Reporter: Greg Harris >Assignee: Linu Shibu >Priority: Trivial > Labels: newbie > > The RemoteLogMetadataSerde#serialize receives a RemoteLogMetadata object, and > has to dispatch to one of four serializers depending on it's type. This is > done by taking the class name of the RemoteLogMetadata and looking it up in > maps to find the corresponding serializer for that class. > This later requires an unchecked cast, because the RemoteLogMetadataTransform > is generic. This is all type-unsafe, and can be replaced with type-safe > if-elseif-else statements that may also be faster than the double-indirect > map lookups. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Assigned] (KAFKA-16356) Remove class-name dispatch in RemoteLogMetadataSerde
[ https://issues.apache.org/jira/browse/KAFKA-16356?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Linu Shibu reassigned KAFKA-16356: -- Assignee: Linu Shibu > Remove class-name dispatch in RemoteLogMetadataSerde > > > Key: KAFKA-16356 > URL: https://issues.apache.org/jira/browse/KAFKA-16356 > Project: Kafka > Issue Type: Task > Components: Tiered-Storage >Affects Versions: 3.7.0 >Reporter: Greg Harris >Assignee: Linu Shibu >Priority: Trivial > Labels: newbie > > The RemoteLogMetadataSerde#serialize receives a RemoteLogMetadata object, and > has to dispatch to one of four serializers depending on it's type. This is > done by taking the class name of the RemoteLogMetadata and looking it up in > maps to find the corresponding serializer for that class. > This later requires an unchecked cast, because the RemoteLogMetadataTransform > is generic. This is all type-unsafe, and can be replaced with type-safe > if-elseif-else statements that may also be faster than the double-indirect > map lookups. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Comment Edited] (KAFKA-16356) Remove class-name dispatch in RemoteLogMetadataSerde
[ https://issues.apache.org/jira/browse/KAFKA-16356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17824971#comment-17824971 ] Linu Shibu edited comment on KAFKA-16356 at 3/9/24 5:30 PM: Can I assign this to myself and work on it [~gharris1727] ? was (Author: JIRAUSER304485): Can I assign this to myself and work on it? > Remove class-name dispatch in RemoteLogMetadataSerde > > > Key: KAFKA-16356 > URL: https://issues.apache.org/jira/browse/KAFKA-16356 > Project: Kafka > Issue Type: Task > Components: Tiered-Storage >Affects Versions: 3.7.0 >Reporter: Greg Harris >Priority: Trivial > Labels: newbie > > The RemoteLogMetadataSerde#serialize receives a RemoteLogMetadata object, and > has to dispatch to one of four serializers depending on it's type. This is > done by taking the class name of the RemoteLogMetadata and looking it up in > maps to find the corresponding serializer for that class. > This later requires an unchecked cast, because the RemoteLogMetadataTransform > is generic. This is all type-unsafe, and can be replaced with type-safe > if-elseif-else statements that may also be faster than the double-indirect > map lookups. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Commented] (KAFKA-16356) Remove class-name dispatch in RemoteLogMetadataSerde
[ https://issues.apache.org/jira/browse/KAFKA-16356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17824971#comment-17824971 ] Linu Shibu commented on KAFKA-16356: Can I assign this to myself and work on it? > Remove class-name dispatch in RemoteLogMetadataSerde > > > Key: KAFKA-16356 > URL: https://issues.apache.org/jira/browse/KAFKA-16356 > Project: Kafka > Issue Type: Task > Components: Tiered-Storage >Affects Versions: 3.7.0 >Reporter: Greg Harris >Priority: Trivial > Labels: newbie > > The RemoteLogMetadataSerde#serialize receives a RemoteLogMetadata object, and > has to dispatch to one of four serializers depending on it's type. This is > done by taking the class name of the RemoteLogMetadata and looking it up in > maps to find the corresponding serializer for that class. > This later requires an unchecked cast, because the RemoteLogMetadataTransform > is generic. This is all type-unsafe, and can be replaced with type-safe > if-elseif-else statements that may also be faster than the double-indirect > map lookups. -- This message was sent by Atlassian Jira (v8.20.10#820010)