cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bryan <br...@appssavvy.com>
Subject Diagnosing a row caching problem
Date Thu, 01 Nov 2012 16:01:05 GMT

Having a problem diagnosing an issue with row caching. It seems like row caching is not working
(very few items stored), despite it being enabled, using JNA, and the key cache being super
hot.  I assume I'm missing something obvious, but I would expect to have more items stored
in the row cache, even after updates. Is there something in the logs I should look for? Are
writes invalidating the cache?

Here are the stats:

keys: 100 bytes, values: 200 - 400 bytes, with read and write pattern

Single node, Cassandra 1.1.5, test node

CFSTATS

Keyspace: <omitted>
	Read Count: 15967248
	Read Latency: 0.4956699828924809 ms.
	Write Count: 15967240
	Write Latency: 0.027375880803445052 ms.
	Pending Tasks: 0
		Column Family:  <omitted>
		SSTable count: 75
		Space used (live): 705364536
		Space used (total): 705364536
		Number of Keys (estimate): 2591104
		Memtable Columns Count: 267840
		Memtable Data Size: 129949276
		Memtable Switch Count: 192
		Read Count: 15967248
		Read Latency: NaN ms.
		Write Count: 15967240
		Write Latency: NaN ms.
		Pending Tasks: 0
		Bloom Filter False Postives: 281
		Bloom Filter False Ratio: 0.00000
		Bloom Filter Space Used: 3986944
		Compacted row minimum size: 311
		Compacted row maximum size: 642
		Compacted row mean size: 419


 INFO

Load             : 676.5 MB
Heap Memory (MB) : 1476.10 / 2925.00
Key Cache    : size 104857584 (bytes), capacity 104857584 (bytes), 6146639 hits, 14680951
requests, 1.000 recent hit rate, 14400 save period in seconds
Row Cache  : size 0 (bytes), capacity 209715200 (bytes), 47 hits, 14400100 requests, NaN recent
hit rate, 87000 save period in seconds


Mime
View raw message