[
https://issues.apache.org/jira/browse/IGNITE-27149?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ivan Bessonov updated IGNITE-27149:
-----------------------------------
Description:
What to remove:
* "Copier has loaded a portion of multi-versioned data"
We don't need to log each request, it pollutes the logs.
What to add/change:
* Clearly log every phase change of the operation. Current phases are:
** Loading metadata
** Waiting for catalog
** Preparing storages
** Loading MV data
** Loading TX data
** Finalizing the operation
* In addition to that, at the end of operation we should print the statistics:
** How long it took
** The duration of each phase
** The number of entries / volume of each phase, whatever information we have
available
*** This could be the number of rows, the number of row versions, the number
of transactions, the number of batches, etc.
** Anything else that's useful, be creative
was:
What to remove:
* "Copier has loaded a portion of multi-versioned data"
We don't need to log each request, it pollutes the logs.
What to add/change:
* Clearly log every phase change of the operation. Current phases are:
** Loading metadata
** Waiting for catalog
** Preparing storages
** Loading MV data
** Loading TX data
** Finalizing the operation
* In addition to that, at the end of operation we should print the statistics:
** How long it took
** The duration of each phase
** The number of entries / volume of each phase, whatever information we have
available
*** This could be the number of rows, the number of row versions, the number
of transactions, etc.
** Anything else that's useful, be creative
> Improve logging in IncomingSnapshotCopier
> -----------------------------------------
>
> Key: IGNITE-27149
> URL: https://issues.apache.org/jira/browse/IGNITE-27149
> Project: Ignite
> Issue Type: Improvement
> Components: raft snapshots ai3
> Reporter: Ivan Bessonov
> Priority: Major
> Labels: ignite-3
>
> What to remove:
> * "Copier has loaded a portion of multi-versioned data"
> We don't need to log each request, it pollutes the logs.
> What to add/change:
> * Clearly log every phase change of the operation. Current phases are:
> ** Loading metadata
> ** Waiting for catalog
> ** Preparing storages
> ** Loading MV data
> ** Loading TX data
> ** Finalizing the operation
> * In addition to that, at the end of operation we should print the
> statistics:
> ** How long it took
> ** The duration of each phase
> ** The number of entries / volume of each phase, whatever information we
> have available
> *** This could be the number of rows, the number of row versions, the number
> of transactions, the number of batches, etc.
> ** Anything else that's useful, be creative
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
