You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ignite.apache.org by "К. Алексей" <ka...@yandex.ru> on 2019/09/20 09:20:20 UTC
Unable to understand Primary/Backup entries (Off-heap, replicated cache)
Hi,
I have a cluster with three baseline nodes, one data region and one replicated
cache. When I look into ignitevisor cache info (cache -a) I see primary and
backup entries in off-heap in section "Nodes for: ***", Column Size (Primary /
Backup), Off-Heap row:
| Node ID8(@), IP | CPUs | Heap Used | CPU Load | Up Time | Size (Primary /
Backup) | Hi/Mi/Rd/Wr |
+================================================================|
3B116797(@n4) | 4 | 59.75 % | 0.33 % | 01:29:33.654 | Total: 3023 (1767 /
1256) | Hi: 13 |
| | | | | | Heap: 1103 (1103 / <n/a>) | Mi: 1584 |
| | | | | | Off-Heap: 1920 (664 / 1256) | Rd: 1597 |
| | | | | | Off-Heap Memory: <n/a> | Wr: 5379 |
+-------------------------------------------------------------------------------------------------------------------------+
Cache config:
<bean class="org.apache.ignite.configuration.CacheConfiguration">
<property name="name" value="sessions"/>
<property name="cacheMode" value="REPLICATED"/>
<property name="onheapCacheEnabled" value="true"/>
<property name="statisticsEnabled" value="true"/>
<property name="eagerTtl" value="true"/>
<property name="evictionPolicy">
<bean class="org.apache.ignite.cache.eviction.lru.LruEvictionPolicy">
<property name="maxSize" value="3000"/>
</bean>
</property>
<property name="expiryPolicyFactory">
<bean class="javax.cache.expiry.TouchedExpiryPolicy" factory-
method="factoryOf">
<constructor-arg>
<bean class="javax.cache.expiry.Duration">
<constructor-arg value="MINUTES"/>
<constructor-arg value="60"/>
</bean>
</constructor-arg>
</bean>
</property>
</bean>
Data region config:
<property name="dataStorageConfiguration">
<bean class="org.apache.ignite.configuration.DataStorageConfiguration">
<property name="defaultDataRegionConfiguration">
<bean class="org.apache.ignite.configuration.DataRegionConfiguration">
<property name="metricsEnabled" value="true"/>
<property name="persistenceEnabled" value="true"/>
<property name="initialSize" value="#{256L * 1024 * 1024}"/>
<property name="maxSize" value="#{1280L * 1024 * 1024}"/>
<property name="pageEvictionMode" value="RANDOM_LRU"/>
</bean>
</property>
</bean>
</property>
Few things that I can't understand:
1) Is it normal that cache have some Backup entries in REPLICATED mode?
2) Can I disable or clean this backup entries?
The primary problem is that my baseline nodes has different count of Off-heap
entries: 431, 336, 1020. And third node (1020) always have some constant
number of Off-heap entries (backup as I think) in case that eviction and
expire policies are set.
Re: Unable to understand Primary/Backup entries (Off-heap, replicated cache)
Posted by Igor Belyakov <ig...@gmail.com>.
Hi,
1. Yes, it's normal, since in REPLICATED mode you have primary copy of
every key and backup copies on all of the nodes.
More information here:
https://apacheignite.readme.io/docs/cache-modes#section-replicated-mode
2. In case you don't want to have backup copies then you should choose
PARTITIONED mode with BACKUP=0, but in this mode it will be not possible to
restore the data in case of data lost on one of the nodes.
https://apacheignite.readme.io/docs/cache-modes#section-replicated-mode
3. According to the provided information, I see that your node has 1920
entries in total (664 primary and 1256 backups for 2 other nodes).
Regards,
Igor
On Fri, Sep 20, 2019 at 12:20 PM К. Алексей <ka...@yandex.ru> wrote:
> Hi,
>
> I have a cluster with three baseline nodes, one data region and one
> replicated cache. When I look into ignitevisor cache info (cache -a) I see
> primary and backup entries in off-heap in section "Nodes for: ***", Column
> Size (Primary / Backup), Off-Heap row:
>
> | Node ID8(@), IP | CPUs | Heap Used | CPU Load | Up Time
> | Size (Primary / Backup) | Hi/Mi/Rd/Wr |
> +================================================================|
> 3B116797(@n4) | 4 | 59.75 % | 0.33 % | 01:29:33.654 | Total: 3023
> (1767 / 1256) | Hi: 13 |
> | | | | |
> | Heap: 1103 (1103 / <n/a>) | Mi: 1584 |
> | | | | |
> | Off-Heap: 1920 (664 / 1256) | Rd: 1597 |
> | | | | |
> | Off-Heap Memory: <n/a> | Wr: 5379 |
>
> +-------------------------------------------------------------------------------------------------------------------------+
>
> Cache config:
> <bean class="org.apache.ignite.configuration.CacheConfiguration">
> <property name="name" value="sessions"/>
> <property name="cacheMode" value="REPLICATED"/>
> <property name="onheapCacheEnabled" value="true"/>
> <property name="statisticsEnabled" value="true"/>
> <property name="eagerTtl" value="true"/>
> <property name="evictionPolicy">
> <bean
> class="org.apache.ignite.cache.eviction.lru.LruEvictionPolicy">
> <property name="maxSize" value="3000"/>
> </bean>
> </property>
> <property name="expiryPolicyFactory">
> <bean class="javax.cache.expiry.TouchedExpiryPolicy"
> factory-method="factoryOf">
> <constructor-arg>
> <bean class="javax.cache.expiry.Duration">
> <constructor-arg value="MINUTES"/>
> <constructor-arg value="60"/>
> </bean>
> </constructor-arg>
> </bean>
> </property>
> </bean>
>
> Data region config:
> <property name="dataStorageConfiguration">
> <bean class="org.apache.ignite.configuration.DataStorageConfiguration">
> <property name="defaultDataRegionConfiguration">
> <bean
> class="org.apache.ignite.configuration.DataRegionConfiguration">
> <property name="metricsEnabled" value="true"/>
> <property name="persistenceEnabled" value="true"/>
> <property name="initialSize" value="#{256L * 1024 *
> 1024}"/>
> <property name="maxSize" value="#{1280L * 1024 * 1024}"/>
> <property name="pageEvictionMode" value="RANDOM_LRU"/>
> </bean>
> </property>
> </bean>
> </property>
>
> Few things that I can't understand:
> 1) Is it normal that cache have some Backup entries in REPLICATED mode?
> 2) Can I disable or clean this backup entries?
>
> The primary problem is that my baseline nodes has different count of
> Off-heap entries: 431, 336, 1020. And third node (1020) always have some
> constant number of Off-heap entries (backup as I think) in case that
> eviction and expire policies are set.
>