[
https://issues.apache.org/jira/browse/IGNITE-29018?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ignite TC Bot updated IGNITE-29018:
-----------------------------------
Description:
Document the Apache Ignite 2 cache dump reader utility/API as one coherent
user-facing topic.
This follow-up intentionally covers more than the original IGNITE-27716 log
wording. The current public docs did not have a dedicated cache dump reader
page, so a narrow paragraph in the snapshots restore section would be
misleading.
Source scope:
* IGNITE-19950 - cache dumps
* IGNITE-20472 - API to consume cache dumps
* IGNITE-20601 - custom dump location handling
* IGNITE-20614 - snapshotTransferRate used for dump creation rate limiting
* IGNITE-20708 - cache group filter for DumpReader
* IGNITE-20722 - skipCopies for DumpReader
* IGNITE-22020 - keepRaw parameter for DumpReaderConfiguration
* IGNITE-25554 - cache names filter for DumpReader
* IGNITE-26729 - non-volatile data structures in cache dump
* IGNITE-27716 - cache group/cache name in DumpReader partition logs
* IGNITE-27842 - DumpReader consumer/API no longer marked experimental
User-facing documentation should explain:
* how to create a cache dump through IgniteSnapshot.createDump(...);
* how to read a dump with DumpReader, DumpReaderConfiguration, DumpConsumer,
and DumpEntry;
* the consumer lifecycle and the fact that onPartition(...) can be called
concurrently;
* reader options: threadCount, timeout, failFast, keepBinary, keepRaw,
groupNames, cacheNames, skipCopies, and encryptionSpi;
* that snapshotTransferRate also limits cache dump creation disk write rate;
* how to interpret DumpReader partition log messages, including the grp field.
Draft PR: https://github.com/apache/ignite/pull/13521
was:
IGNITE-27716 added cache group names to DumpReader log messages.
The implementation PR is https://github.com/apache/ignite/pull/12840, merged as
ffaea42d2e9d8fbac8dd96dcc50f99c14028ec22.
Current user-facing snapshot documentation explains snapshot restore and cache
group selection, but it does not mention how to interpret DumpReader log
messages during dump processing. A small documentation note should say that
DumpReader log messages such as "Skip copy partition", "Skip partition due to
previous error", "Consuming partition", and "Error consuming partition" now
print grp=<cache-or-group-name>, which helps map dump processing output back to
the affected cache group.
Summary: [doc] Document cache dump reader utility and API (was: [doc]
Document DumpReader cache group names in logs)
> [doc] Document cache dump reader utility and API
> ------------------------------------------------
>
> Key: IGNITE-29018
> URL: https://issues.apache.org/jira/browse/IGNITE-29018
> Project: Ignite
> Issue Type: Task
> Components: documentation
> Reporter: Ignite TC Bot
> Priority: Major
> Labels: ignite-2
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Document the Apache Ignite 2 cache dump reader utility/API as one coherent
> user-facing topic.
> This follow-up intentionally covers more than the original IGNITE-27716 log
> wording. The current public docs did not have a dedicated cache dump reader
> page, so a narrow paragraph in the snapshots restore section would be
> misleading.
> Source scope:
> * IGNITE-19950 - cache dumps
> * IGNITE-20472 - API to consume cache dumps
> * IGNITE-20601 - custom dump location handling
> * IGNITE-20614 - snapshotTransferRate used for dump creation rate limiting
> * IGNITE-20708 - cache group filter for DumpReader
> * IGNITE-20722 - skipCopies for DumpReader
> * IGNITE-22020 - keepRaw parameter for DumpReaderConfiguration
> * IGNITE-25554 - cache names filter for DumpReader
> * IGNITE-26729 - non-volatile data structures in cache dump
> * IGNITE-27716 - cache group/cache name in DumpReader partition logs
> * IGNITE-27842 - DumpReader consumer/API no longer marked experimental
> User-facing documentation should explain:
> * how to create a cache dump through IgniteSnapshot.createDump(...);
> * how to read a dump with DumpReader, DumpReaderConfiguration, DumpConsumer,
> and DumpEntry;
> * the consumer lifecycle and the fact that onPartition(...) can be called
> concurrently;
> * reader options: threadCount, timeout, failFast, keepBinary, keepRaw,
> groupNames, cacheNames, skipCopies, and encryptionSpi;
> * that snapshotTransferRate also limits cache dump creation disk write rate;
> * how to interpret DumpReader partition log messages, including the grp field.
> Draft PR: https://github.com/apache/ignite/pull/13521
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
