[ https://issues.apache.org/jira/browse/CASSANDRA-11383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15213161#comment-15213161 ]
DOAN DuyHai commented on CASSANDRA-11383: ----------------------------------------- *Test case 2* Ok [~xedin], I've found something really *fishy* On the seed node, in the logs: {noformat} INFO [GossipStage:1] 2016-03-26 16:35:22,020 Gossiper.java:1009 - InetAddress /46.105.104.166 is now DOWN INFO [GossipStage:1] 2016-03-26 16:36:18,474 Gossiper.java:1009 - InetAddress /5.39.73.149 is now DOWN INFO [GossipStage:1] 2016-03-26 16:36:27,285 Gossiper.java:1009 - InetAddress /5.39.72.230 is now DOWN {noformat} Inside system.log of *46.105.104.166*, around *2016-03-26 16:35:22,020*: {noformat} ERROR [SASI-General:2] 2016-03-26 16:35:22,017 CassandraDaemon.java:195 - Exception in thread Thread[SASI-General:2,5,main] org.apache.cassandra.io.FSWriteError: java.io.IOException: Term - '-9140913758573873094' belongs to more than 5 keys in SPARSE mode, which is not allowed. at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:301) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:244) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:227) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.PerSSTableIndexWriter$Index.lambda$scheduleSegmentFlush$262(PerSSTableIndexWriter.java:262) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at java.util.concurrent.FutureTask.run(FutureTask.java:266) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_74] at java.lang.Thread.run(Thread.java:745) [na:1.8.0_74] Caused by: java.io.IOException: Term - '-9140913758573873094' belongs to more than 5 keys in SPARSE mode, which is not allowed. {noformat} Inside system.log of *5.39.73.149*, around *2016-03-26 16:36:18,474*: {noformat} ERROR [SASI-General:2] 2016-03-26 16:36:18,473 CassandraDaemon.java:195 - Exception in thread Thread[SASI-General:2,5,main] org.apache.cassandra.io.FSWriteError: java.io.IOException: Term - '-8889829184905116393' belongs to more than 5 keys in SPARSE mode, which is not allowed. at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:301) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:244) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:227) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.PerSSTableIndexWriter$Index.lambda$scheduleSegmentFlush$262(PerSSTableIndexWriter.java:262) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at java.util.concurrent.FutureTask.run(FutureTask.java:266) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_74] at java.lang.Thread.run(Thread.java:745) [na:1.8.0_74] Caused by: java.io.IOException: Term - '-8889829184905116393' belongs to more than 5 keys in SPARSE mode, which is not allowed. {noformat} Inside system.log of *5.39.72.230*, around *2016-03-26 16:36:27,285*: {noformat} ERROR [SASI-General:2] 2016-03-26 16:36:27,281 CassandraDaemon.java:195 - Exception in thread Thread[SASI-General:2,5,main] org.apache.cassandra.io.FSWriteError: java.io.IOException: Term - '-8574128631854799776' belongs to more than 5 keys in SPARSE mode, which is not allowed. at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:301) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:244) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.OnDiskIndexBuilder.finish(OnDiskIndexBuilder.java:227) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at org.apache.cassandra.index.sasi.disk.PerSSTableIndexWriter$Index.lambda$scheduleSegmentFlush$262(PerSSTableIndexWriter.java:262) ~[apache-cassandra-3.5-SNAPSHOT.jar:3.5-SNAPSHOT] at java.util.concurrent.FutureTask.run(FutureTask.java:266) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) ~[na:1.8.0_74] at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_74] at java.lang.Thread.run(Thread.java:745) [na:1.8.0_74] Caused by: java.io.IOException: Term - '-8574128631854799776' belongs to more than 5 keys in SPARSE mode, which is not allowed. {noformat} I checked a couple of other nodes and the same error pattern is seen ... Look like the SASI exception raised on {{SPARSE}} index just kill the gossip. I don't see how, maybe the exception just bubble up the stack and kill the gossip stage ? > Avoid index segment stitching in RAM which lead to OOM on big SSTable files > ---------------------------------------------------------------------------- > > Key: CASSANDRA-11383 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11383 > Project: Cassandra > Issue Type: Bug > Components: CQL > Environment: C* 3.4 > Reporter: DOAN DuyHai > Assignee: Jordan West > Labels: sasi > Fix For: 3.5 > > Attachments: CASSANDRA-11383.patch, > SASI_Index_build_LCS_1G_Max_SSTable_Size_logs.tar.gz, > new_system_log_CMS_8GB_OOM.log, system.log_sasi_build_oom > > > 13 bare metal machines > - 6 cores CPU (12 HT) > - 64Gb RAM > - 4 SSD in RAID0 > JVM settings: > - G1 GC > - Xms32G, Xmx32G > Data set: > - ≈ 100Gb/per node > - 1.3 Tb cluster-wide > - ≈ 20Gb for all SASI indices > C* settings: > - concurrent_compactors: 1 > - compaction_throughput_mb_per_sec: 256 > - memtable_heap_space_in_mb: 2048 > - memtable_offheap_space_in_mb: 2048 > I created 9 SASI indices > - 8 indices with text field, NonTokenizingAnalyser, PREFIX mode, > case-insensitive > - 1 index with numeric field, SPARSE mode > After a while, the nodes just gone OOM. > I attach log files. You can see a lot of GC happening while index segments > are flush to disk. At some point the node OOM ... > /cc [~xedin] -- This message was sent by Atlassian JIRA (v6.3.4#6332)