[
https://issues.apache.org/jira/browse/CASSANDRA-11383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15216917#comment-15216917
]
Jordan West commented on CASSANDRA-11383:
-----------------------------------------
The docs look pretty comprehensive. Thanks! I'll make a more detailed pass
through them when I get a chance. I think the only thing we would like to
clarify, based on the discussion in this ticket, is when to choose {{SPARSE}}
over {{PREFIX}} for numerical data. My last comment
(https://issues.apache.org/jira/browse/CASSANDRA-11383?focusedCommentId=15216337&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15216337)
mentions a way to do that.
Otherwise, specific to {{SPARSE}} the only recommendation I have is that the
{{SPARSE}} example on the "CREATE CUSTOM INDEX (SASI)" page
(https://docs.datastax.com/en/cql/3.3/cql/cql_reference/refCreateSASIIndex.html)
uses {{age}}, which typically would not be a good candidate for a {{SPARSE}}
index (the answer to question number 2 in my linked comment would be: no, there
are not millions of ages with each term having a small number of matching
keys).
> 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)
