You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Maxim Muzafarov (Jira)" <ji...@apache.org> on 2023/03/09 13:58:00 UTC
[jira] [Updated] (IGNITE-15287) Yardstick's -cwd params does not clean persistence and wal directories
[ https://issues.apache.org/jira/browse/IGNITE-15287?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Maxim Muzafarov updated IGNITE-15287:
-------------------------------------
Fix Version/s: (was: 2.10)
> Yardstick's -cwd params does not clean persistence and wal directories
> ----------------------------------------------------------------------
>
> Key: IGNITE-15287
> URL: https://issues.apache.org/jira/browse/IGNITE-15287
> Project: Ignite
> Issue Type: Bug
> Environment: -
> Reporter: Anton Vinogradov
> Priority: Major
>
> The current code at IgniteNode.java is
> {noformat}
> if (args.cleanWorkDirectory() && !clientMode)
> FileUtils.cleanDirectory(U.workDirectory(c.getWorkDirectory(), c.getIgniteHome()));
> {noformat}
> and we should append {{dataStorageConfiguration#storagePath}} and {{dataStorageConfiguration#walPath}} clean-up as well.
> Why it's important:
> For example, hawing following configuration
> {noformat}
> <bean id="dataStorageConfiguration" class="org.apache.ignite.configuration.DataStorageConfiguration">
> <property name="defaultDataRegionConfiguration">
> <ref bean="dataRegionConfiguration"/>
> </property>
> <!-- Sets property that defines behavior of wal fsync. -->
> <property name="walMode" value="LOG_ONLY"/>
> <property name="pageSize" value="#{4 * 1024}"/>
> <!-- Sets a total number of checkpoints to keep in the WAL history. -->
> <property name="walHistorySize" value="2147483647"/>
> <!-- Sets size of a WAL segment. -->
> <property name="walSegmentSize" value="#{1024*1024*1024}"/>
> <property name="storagePath" value="/mnt/ssd/persistence"/>
> <property name="walPath" value="/mnt/ssd/wal"/>
> </bean>
> {noformat}
> We will use same persistence each time we running benchmarks, this means we'll gain different numbers each time :(
> {{/mnt/ssd/wal}} will grow until possible, and {{/mnt/ssd/persistence}} will be different before each benchmark start.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)