[ https://issues.apache.org/jira/browse/CASSANDRA-13937?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16192062#comment-16192062 ]
Andrey Lataev commented on CASSANDRA-13937: ------------------------------------------- LZ4 compression used: {code:java} sys@cqlsh> DESC TABLE p00smevauditbody.messagelogbody20171004; CREATE TABLE p00smevauditbody.messagelogbody20171004 ( d timestamp, mid text, mt text, rec_msg text, rec_sid text, sd_msg text, PRIMARY KEY (d, mid, mt) ) WITH CLUSTERING ORDER BY (mid ASC, mt ASC) AND bloom_filter_fp_chance = 0.1 AND caching = {'keys': 'ALL', 'rows_per_partition': 'NONE'} AND comment = '' AND compaction = {'class': 'org.apache.cassandra.db.compaction.LeveledCompactionStrategy', 'sstable_size_in_mb': '160'} AND compression = {'chunk_length_in_kb': '64', 'class': 'org.apache.cassandra.io.compress.LZ4Compressor'} AND crc_check_chance = 1.0 AND dclocal_read_repair_chance = 0.1 AND default_time_to_live = 0 AND gc_grace_seconds = 864000 AND max_index_interval = 2048 AND memtable_flush_period_in_ms = 0 AND min_index_interval = 128 AND read_repair_chance = 0.0 AND speculative_retry = '99PERCENTILE'; {code} > Cassandra node's startup time increased after increase count of big tables > --------------------------------------------------------------------------- > > Key: CASSANDRA-13937 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13937 > Project: Cassandra > Issue Type: Bug > Components: Core > Environment: RHEL 7.3 > JDK HotSpot 1.8.0_121-b13 > cassandra-3.11 cluster with 43 nodes in 9 datacenters > 8vCPU, 100 GB RAM > Reporter: Andrey Lataev > Attachments: cassandra.zip, debug.zip > > > In startup time Cassandra spends a long time on read some big Columnfamilies. > For example, in debug.log: > {code:java} > grep SSTableReader.java:506 /var/log/cassandra/debug.log > <...> > DEBUG [SSTableBatchOpen:3] 2017-10-04 22:40:05,297 SSTableReader.java:506 - > Opening > /egov/data/cassandra/datafiles1/p00smevauditbody/messagelogbody20171003-b341cc709c7511e7b1cfed1e90eb03dc/mc-45242-big > (19.280MiB) > DEBUG [SSTableBatchOpen:5] 2017-10-04 22:42:14,188 SSTableReader.java:506 - > Opening > /egov/data/cassandra/datafiles1/p00smevauditbody/messagelogbody20171004-f82225509d3e11e7b1cfed1e90eb03dc/mc-49002-big > (10.607MiB) > <...> > DEBUG [SSTableBatchOpen:4] 2017-10-04 22:42:19,792 SSTableReader.java:506 - > Opening > /egov/data/cassandra/datafiles1/p00smevauditbody/messagelogbody20171004-f82225509d3e11e7b1cfed1e90eb03dc/mc-47907-big > (128.172MiB) > DEBUG [SSTableBatchOpen:1] 2017-10-04 22:44:23,560 SSTableReader.java:506 - > Opening > /egov/data/cassandra/datafiles1/pk4smevauditbody/messagelogbody20170324-f918bfa0107b11e7adfc2d0b45a372ac/mc-4-big > (96.310MiB) > <..> > {code} > SSTableReader.java:506 spent ~ 2 min per every big table in p00smevauditbody > keyspace. > I was planned too keep similar tables for the full month... > So it seems like Cassandra will need more then 1h on startup... > Does it available to speed up SSTableBatchOpen ? -- This message was sent by Atlassian JIRA (v6.4.14#64029) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org