Thanks Ali! I use a 13 months TTL on this table. I guess I need to remodel this table. And I'll definitely try this tool.
On Tue, Apr 3, 2018 at 1:28 AM, Ali Hubail <ali.hub...@petrolink.com> wrote: > system.log should show you some warnings about wide rows. Do a grep on > system.log for 'Writing large partition' The message could be different for > the c* version you're using though. Plus, this doesn't show you all of the > large partitions. > > There is a nice tool that analyzes sstables and can show the large > partitions: > https://github.com/tolbertam/sstable-tools > > > By "how to deal with them?" it depends. If you don't need those > partitions then you can delete them. You can also use TTL if it fits you or > remodel your table to only hold upto 100k rows or 100mb per partition > (whichever comes first). If you're going to remodel the table, aim for much > less than 100k/100mb per partition. > > *Ali Hubail* > > Confidentiality warning: This message and any attachments are intended > only for the persons to whom this message is addressed, are confidential, > and may be privileged. If you are not the intended recipient, you are > hereby notified that any review, retransmission, conversion to hard copy, > copying, modification, circulation or other use of this message and any > attachments is strictly prohibited. If you receive this message in error, > please notify the sender immediately by return email, and delete this > message and any attachments from your system. Petrolink International > Limited its subsidiaries, holding companies and affiliates disclaims all > responsibility from and accepts no liability whatsoever for the > consequences of any unauthorized person acting, or refraining from acting, > on any information contained in this message. For security purposes, staff > training, to assist in resolving complaints and to improve our customer > service, email communications may be monitored and telephone calls may be > recorded. > > > *shalom sagges <shalomsag...@gmail.com <shalomsag...@gmail.com>>* > > 04/02/2018 03:57 PM > Please respond to > user@cassandra.apache.org > > To > user@cassandra.apache.org, > > cc > Subject > Large Partitions > > > > > Hi All, > > I ran nodetool cfstats (v2.0.14) on a keyspace and found that there are a > few large partitions. I assume that since "Compacted partition maximum > bytes": 802187438 (~800 MB) and since > "Compacted partition mean bytes": 100465 (~100 KB), it means that most > partitions are in okay size and only a few are large. Am I assuming > correctly? > > If so, can anyone suggest how to find those large partitions and how to > deal with them? (cfstats output below) > > Thanks! > > > nodetool cfstats keyspace1; > > Table: table1 > SSTable count: 16 > Space used (live), bytes: 453844035587 > Space used (total), bytes: 453844035587 > Off heap memory used (total), bytes: 440787635 > SSTable Compression Ratio: 0.17417149031966575 > Number of keys (estimate): 33651200 > Memtable cell count: 27966 > Memtable data size, bytes: 41698140 > Memtable switch count: 199727 > Local read count: 86494530 > Local read latency: 2.646 ms > Local write count: 247712138 > Local write latency: 0.030 ms > Pending tasks: 0 > Bloom filter false positives: 2182242 > Bloom filter false ratio: 0.02251 > Bloom filter space used, bytes: 53135136 > Bloom filter off heap memory used, bytes: 53135008 > Index summary off heap memory used, bytes: 11560419 > Compression metadata off heap memory used, bytes: 376092208 > Compacted partition minimum bytes: 373 > > * Compacted partition maximum bytes: 802187438 Compacted partition mean > bytes: 100465* > Average live cells per slice (last five minutes): 37.0 > Average tombstones per slice (last five minutes): 0.0 > >
