I found this topic How do i exclude everything but text/html from a heritrix crawl?
I have changed bean to this
<property name="shouldProcessRule">
<bean class="org.archive.modules.deciderules.ContentTypeMatchesRegexDecideRule">
<property name="decision" value="ACCEPT" />
<property name="regex" value="^application/pdf.*"/>
</bean>
</property>
</bean>
But heritrix still saves every file to mirror dir.