Illustration Image

2/17/2023

Reading time:10

ic-tools for Apache Cassandra SSTables

Placeholder

This resource is based on an article originally published here.

Table of Contents

Overview

Instaclustr has developed a number of useful tools to assist with diagnosing issues in a cluster. For users of Instaclustr’s Managed Service, our Technical Operations team will run these as needed when working with you to help diagnose issues. The tools are available on a supported basis for our enterprise support customers and on an unsupported basis for the general community (although we’ll probably answer questions on the C* user email list).

These tools supplement the information available from the nodetool utility that is part of core Apache Cassandra. Whereas nodetool tends to report based on summary statistics maintained as Cassandra services operate, ic-tools directly read Cassandra’s data files when executed to report more detailed and accurate statistics.

As such, executing the tools can result in a large amount of data being read which can potentially impact the performance of a node where they are being executed. The two most data heavy tools (ic-cfstats and ic-purge) provide rate limiting functions to reduce the impact. However, users are advised to execute care when using these tools in a live cluster.

These tools are version-specific and you must use the corresponding ic-tools version for your Cassandra version. We have provided pre-built jars for all versions of Cassandra at the bottom of this page.

The source code is published on GitHub.

Command Description
ic-summary Summary information about all column families including how much of the data is repaired
ic-sstables Print out metadata for SSTables the belong to a column family
ic-pstats Partition size statistics for a column family
ic-cfstats Detailed statistics about cells in a column family
ic-purge Statistics about reclaimable data for a column family

(We’ve generally used the old-school C* term ‘column family’. It is synonymous with ‘table’ in modern C* versions.)

ic-summary

Provides summary information about all column families. Useful for finding the largest column families and how much data has been repaired by incremental repairs.

Usage

ic-summary

Output

Column Description
Keyspace Keyspace the column family belongs to
Column Family Name of column family
SSTables Number of SSTables on this node for the column family
Disk Size Compressed size on disk for this node
Data Size Uncompressed size of the data for this node
Last Repaired Time of the last incremental repair
Repair % Percentage of data marked as repaired by incremental repair

ic-sstables

Print out SSTable metadata for a column family. Useful in helping to tune compaction settings.

Usage

ic-sstables <keyspace> <column-family>

Output

Column Description
SSTable Data.db filename of SSTable
Disk Size Size of SSTable on disk
Total Size Uncompressed size of data contained in the SSTable
Min Timestamp Minimum cell timestamp contained in the SSTable
Max Timestamp Maximum cell timestamp contained in the SSTable
Duration The time span between minimum and maximum cell timestamps
Level Leveled Tiered Compaction SSTable level
Keys Number of partition keys
Avg Partition Size Average partition size
Max Partition Size Maximum partition size
Avg Column Count Average number of columns in a partition
Max Column Count Maximum number of columns in a partition
Droppable Estimated droppable tombstones
Repaired At Time when marked as repaired by incremental repair

ic-pstats

Tool for finding largest partitions. Reads the Index.db files so is relatively quick.

Usage

ic-pstats [-n <num>] [-t <snapshot>] [-f <filter>] <keyspace> <column-family>

-h Display help
-b Batch mode. Uses progress indicator that is friendly for running in batch jobs.
-n <num> Number of partitions to display in leaders lists
-t <name> Snapshot to analyse (snapshot name from nodetool listsnapshots). Snapshot is created if none is specified.
-f <files> Comma separated list of Data.db SSTables to filter on

Output

Summary: Summary statistics about partitions

Column Description
Count (Size) Number of partition keys on this node
Total (Size) Total uncompressed size of all partitions on this node
Total (SSTable) Number of SSTables on this node
Minimum (Size) Minimum uncompressed partition size
Minimum (SSTable) Minimum number of SSTables a partition belongs to
Average (Size) Average (mean) uncompressed partition size
Average (SSTable) Average (mean) number of SSTables a partition belongs to
Std dev. (Size) Standard deviation of partition sizes
Std dev. (SSTable) Standard deviation of number of SSTables for a partition
50% (Size) Estimated 50th percentile of partition sizes
50% (SSTable) Estimated 50th percentile of SSTables for a partition
75% (Size) Estimated 75th percentile of partition sizes
75% (SSTable) Estimated 75th percentile of SSTables for a partition
90% (Size) Estimated 90th percentile of partition sizes
90% (SSTable) Estimated 90th percentile of SSTables for a partition
95% (Size) Estimated 95th percentile of partition sizes
95% (SSTable) Estimated 95th percentile of SSTables for a partition
99% (Size) Estimated 99th percentile of partition sizes
99% (SSTable) Estimated 99th percentile of SSTables for a partition
99.9% (Size) Estimated 99.9th percentile of partition sizes
99.9% (SSTable) Estimated 99.9th percentile of SSTables for a partition
Maximum (Size) Maximum uncompressed partition size
Maximum (SSTable) Maximum number of SSTables a partition belongs to

Largest partitions: The top N largest partitions

Column Description
Key The partition key
Size Total uncompressed size of the partition
SSTable Count Number of SSTables that contain the partition

SSTable Leaders: The top N partitions that belong to the most SSTables

Column Description
Key The partition key
SSTable Count Number of SSTables that contain the partition
Size Total uncompressed size of the partition

SSTables: Metadata about SSTables as it relates to partitions.

Column Description
SSTable Data.db filename of SSTable
Size Uncompressed size
Min Timestamp Minimum cell timestamp in the SSTable
Max Timestamp Maximum cell timestamp in the SSTable
Level Leveled Tiered Compaction level of SSTable
Partitions Number of partition keys in the SSTable
Avg Partition Size Average uncompressed partition size in SSTable
Max Partition Size Maximum uncompressed partition size in SSTable

ic-cfstats

Tool for getting detailed cell statistics that can help identify issues with data model.

Usage

ic-cfstats [-r <limit>] [-n <num>] [-t <snapshot>] [-f <filter>] <keyspace> <column-family>

-h Display help
-b Batch mode. Uses progress indicator that is friendly for running in batch jobs.
-r <limit> Limit read throughput to ratelimit MB/s (unlimited by default, 16 is probably a good starting point if you want to limit)
-n <num> Number of partitions to display in leaders lists
-t <name> Snapshot to analyse (snapshot name from nodetool listsnapshots). Snapshot is created if none is specified.
-f <files> Comma separated list of Data.db SSTables to filter on

Output

Summary: Summary statistics about partitions

Column Description
Count (Size) Number of partition keys on this node
Rows (Size) (3.x only) Number of clustering rows
(deleted) (3.x only) Number of clustering row deletions
Total (Size) Total uncompressed size of all partitions on this node
Total (SSTable) Number of SSTables on this node
Minimum (Size) Minimum uncompressed partition size
Minimum (SSTable) Minimum number of SSTables a partition belongs to
Average (Size) Average (mean) uncompressed partition size
Average (SSTable) Average (mean) number of SSTables a partition belongs to
Std dev. (Size) Standard deviation of partition sizes
Std dev. (SSTable) Standard deviation of number of SSTables for a partition
50% (Size) Estimated 50th percentile of partition sizes
50% (SSTable) Estimated 50th percentile of SSTables for a partition
75% (Size) Estimated 75th percentile of partition sizes
75% (SSTable) Estimated 75th percentile of SSTables for a partition
90% (Size) Estimated 90th percentile of partition sizes
90% (SSTable) Estimated 90th percentile of SSTables for a partition
95% (Size) Estimated 95th percentile of partition sizes
95% (SSTable) Estimated 95th percentile of SSTables for a partition
99% (Size) Estimated 99th percentile of partition sizes
99% (SSTable) Estimated 99th percentile of SSTables for a partition
99.9% (Size) Estimated 99.9th percentile of partition sizes
99.9% (SSTable) Estimated 99.9th percentile of SSTables for a partition
Maximum (Size) Maximum uncompressed partition size
Maximum (SSTable) Maximum number of SSTables a partition belongs to

(3.x only) Row Histogram: Histogram of number of rows per partition

Column Description
Percentile Minimum, average, standard deviation (std dev.), percentile, maximum
Count Estimated number of rows per partition for the given percentile

Largest partitions: Partitions with largest uncompressed size

Column Description
Key The partition key
Size Total uncompressed size of the partition
Rows (3.x only) Total number of clustering rows in the partition
(deleted) (3.x only) Number of row deletions in the partition
Tombstones Number of cell or range tombstones
(droppable) Number of tombstones that can be dropped as per gc_grace_seconds
Cells Number of cells in the partition
SSTable Count Number of SSTables that contain the partition

Widest partitions: Partitions with the most cells

Column Description
Key The partition key
Rows (3.x only) Total number of clustering rows in the partition
(deleted) (3.x only) Number of row deletions in the partition
Cells Number of cells in the partition
Tombstones Number of cell or range tombstones
(droppable) Number of tombstones that can be dropped as per gc_grace_seconds
Size Total uncompressed size of the partition
SSTable Count Number of SSTables that contain the partition

(3.x only) Most Deleted Rows: Partitions with the most row deletions

Column Description
Key The partition key
Rows Total number of clustering rows in the partition
(deleted) Number of row deletions in the partition
Size Total uncompressed size of the partition
SSTable Count Number of SSTables that contain the partition

Tombstone Leaders: Partitions with the most tombstones

Column Description
Key The partition key
Tombstones Number of cell or range tombstones
(droppable) Number of tombstones that can be dropped as per gc_grace_seconds
Rows (3.x only) Total number of clustering rows in the partition
Cells Number of cells in the partition
Size Total uncompressed size of the partition
SSTable Count Number of SSTables that contain the partition

SSTable Leaders: Partitions that are in the most SSTables

Column Description
Key The partition key
SSTable Count Number of SSTables that contain the partition
Size Total uncompressed size of the partition
Rows (3.x only) Total number of clustering rows in the partition
Cells Number of cells in the partition
Tombstones Number of cell or range tombstones
(droppable) Number of tombstones that can be dropped as per gc_grace_seconds

SSTables: Metadata about SSTables as it relates to partitions.

Column Description
SSTable Data.db filename of SSTable
Size Uncompressed size
Min Timestamp Minimum cell timestamp in the SSTable
Max Timestamp Maximum cell timestamp in the SSTable
Partitions Number of partitions
(deleted) Number of row level partition deletions
(avg size) Average uncompressed partition size in SSTable
(max size) Maximum uncompressed partition size in SSTable
Rows (3.x only) Total number of clustering rows in SSTable
(deleted) (3.x only) Number of row deletions in SSTable
Cells Number of cells in the SSTable
Tombstones Number of cell or range tombstones in the SSTable
(droppable) Number of tombstones that are droppable according to gc_grace_seconds
(range) Number of range tombstones
Cell Liveness Percentage of live cells. Does not consider tombstones or cell updates shadowing cells. That is it is percentage of non-tombstoned cells to total number of cells.

ic-purge

Usage

ic-purge [-r <limit>] [-n <num>] [-t <snapshot>] [-f <filter>] <keyspace> <column-family>

-h Display help
-b Batch mode. Uses progress indicator that is friendly for running in batch jobs.
-r <limit> Limit read throughput to ratelimit MB/s (unlimited by default, 16 is probably a good starting point if you want to limit)
-n <num> Number of partitions to display in leaders lists
-t <name> Snapshot to analyse. Snapshot is created if none is specified.

Output

Largest reclaimable partitions: Partitions with the largest amount of reclaimable data

Column Description
Key The partition key
Size Total uncompressed size of the partition
Reclaim Reclaimable uncompressed size
Generations SSTable generations the partition belongs to

Downloads

By Instaclustr Support

Related Articles

logo
sstable
cassandra
sstables

Explore Further

cassandra

tools

sstables

Become part of our
growing community!
Welcome to Planet Cassandra, a community for Apache Cassandra®! We're a passionate and dedicated group of users, developers, and enthusiasts who are working together to make Cassandra the best it can be. Whether you're just getting started with Cassandra or you're an experienced user, there's a place for you in our community.
A dinosaur
Planet Cassandra is a service for the Apache Cassandra® user community to share with each other. From tutorials and guides, to discussions and updates, we're here to help you get the most out of Cassandra. Connect with us and become part of our growing community today.
© 2009-2023 The Apache Software Foundation under the terms of the Apache License 2.0. Apache, the Apache feather logo, Apache Cassandra, Cassandra, and the Cassandra logo, are either registered trademarks or trademarks of The Apache Software Foundation. Sponsored by Anant Corporation and Datastax, and Developed by Anant Corporation.

Get Involved with Planet Cassandra!

We believe that the power of the Planet Cassandra community lies in the contributions of its members. Do you have content, articles, videos, or use cases you want to share with the world?