HBase Ganglia Monitoring Configuration

HBase Ganglia Monitor HBase construction

1. Hadoop-metrics2-hbase.properties

*. Sink. Ganglia. Class=org. Apache. Hadoop. Metrics2. Sink. Ganglia. Gangliasink31 *. Sink. Ganglia. Period=TenHBase. Sink. Ganglia. Period=TenHBase. Sink. Ganglia. Servers=239.2. One.:8649HBase. Class=org. Apache. Hadoop. Metrics2. Sink. Ganglia. Gangliasink31hbase. Period=TenHBase. Servers==239.2. One.:8649Jvm. Class=org. Apache. Hadoop. Metrics2. Sink. Ganglia. Gangliasink31jvm. Period=TenJvm. Servers==239.2. One.:8649Rpc. Class=org. Apache. Hadoop. Metrics2. Sink. Ganglia. Gangliasink31rpc. Period=TenRpc. Servers==239.2. One.:8649

Restart the cluster after you have modified the configuration. To be able to see the appropriate monitoring information in the ganglia.
Note:
The IP address in the above configuration item is the IP address specified in the gmond.conf file Mcast_join:

/* Feel free toSpecify asMany Udp_send_channels asLike. Gmond used toOnly support has a single channel */udp_send_channel {#bind_hostname = yes # highly recommended, soon toBedefault. # this option tells Gmond to  UseA source address # that resolves toThe machine ' s hostname. Without # This, the metrics may appear toCome from any #Interface  andThe DNS names associated with# those IPs would be used toCreate the Rrds. Mcast_join =239.2. One.Port =8649TTL =1}

###

HBase Monitoring Items

Hbase.regionserver.blockCacheCount: Number of in-memory cache blocks (block cache).

The number of blocks (block) of Storefiles (Hfiles) in the cache.

Hbase.regionserver.blockCacheEvictedCount: The number of blocks freed from the cache due to the heap limit size exceeded.

Hbase.regionserver.blockCacheFree: The available cache block capacity.

Hbase.regionserver.blockCacheHitCachingRatio: Cache Hit Ratio (0-100) for cached blocks. The read cache hit rate is configured to view the cache (such as cacheblocks=true).

Hbase.regionserver.blockCacheHitCount:StoreFiles (Hfiles) The number of blocks read from the cache.

Hbase.regionserver.blockCacheHitRatio: The hit ratio of the cache block (0-100). Contains all read requests, although the Cacheblocks=false readout is counted as cache loss.

Hbase.regionserver.blockCacheMissCount: The number of blocks that were requested by Storefiles (Hfiles) but not read from the cache.

Hbase.regionserver.blockCacheSize: The capacity of a block cache in memory, specifically the amount of memory occupied by Blockcache.

Hbase.regionserver.compactionQueueSize:Compaction Queue size. The number of stores in the regionserver used for compaction.

The number of queued region waits for the flush operation in the Hbase.regionserver.flushQueueSize:MemStore.

Hbase.regionserver.fsReadLatency_avg_time: File system delay (MS). The average time from the HDFs read operation.

Hbase.regionserver.memstoreSizeMB: Total capacity in megabytes (MB) for all memstore in the current regionserver.

Hbase.regionserver.requests: Total number of Read and write requests. The request is equivalent to a regionserver RPC call, so a get is the equivalent of a request, but even if the caching value of the scan is set to 1000, each "next" call to scan is equivalent to a single request. (for example, not per row). For large data block requests, each hfile corresponds to a request.

The sum of the StoreFile file index sizes in Hbase.regionserver.storefileIndexSizeMB:RegionServer (MB).

Specific monitoring items can be viewed Http://hbase.apache.org/book.html#rs_metrics

Copyright notice: This article Bo Master original articles, blogs, without consent may not be reproduced.

HBase Ganglia Monitoring Configuration