Types of compaction

The concept of compaction is used for different kinds of operations in Cassandra, the common thing about these operations is that it takes one or more sstables and output new sstables. The types of compactions are;

Minor compaction
triggered automatically in Cassandra.
Major compaction
a user executes a compaction over all sstables on the node.
User defined compaction
a user triggers a compaction on a given set of sstables.
try to fix any broken sstables. This can actually remove valid data if that data is corrupted, if that happens you will need to run a full repair on the node.
upgrade sstables to the latest version. Run this after upgrading to a new major version.
remove any ranges this node does not own anymore, typically triggered on neighbouring nodes after a node has been bootstrapped since that node will take ownership of some ranges from those nodes.
Secondary index rebuild
rebuild the secondary indexes on the node.
after repair the ranges that were actually repaired are split out of the sstables that existed when repair started.
Sub range compaction
It is possible to only compact a given sub range - this could be useful if you know a token that has been misbehaving - either gathering many updates or many deletes. (nodetool compact -st x -et y) will pick all sstables containing the range between x and y and issue a compaction for those sstables. For STCS this will most likely include all sstables but with LCS it can issue the compaction for a subset of the sstables. With LCS the resulting sstable will end up in L0.

When is a minor compaction triggered?

# When an sstable is added to the node through flushing/streaming etc. # When autocompaction is enabled after being disabled (nodetool enableautocompaction) # When compaction adds new sstables. # A check for new minor compactions every 5 minutes.

Merging sstables

Compaction is about merging sstables, since partitions in sstables are sorted based on the hash of the partition key it is possible to efficiently merge separate sstables. Content of each partition is also sorted so each partition can be merged efficiently.

Tombstones and Garbage Collection (GC) Grace

Why Tombstones

When a delete request is received by Cassandra it does not actually remove the data from the underlying store. Instead it writes a special piece of data known as a tombstone. The Tombstone represents the delete and causes all values which occurred before the tombstone to not appear in queries to the database. This approach is used instead of removing values because of the distributed nature of Cassandra.

Deletes without tombstones

Imagine a three node cluster which has the value [A] replicated to every node.:

[A], [A], [A]

If one of the nodes fails and and our delete operation only removes existing values we can end up with a cluster that looks like:

[], [], [A]

Then a repair operation would replace the value of [A] back onto the two nodes which are missing the value.:

[A], [A], [A]

This would cause our data to be resurrected even though it had been deleted.

Deletes with Tombstones

Starting again with a three node cluster which has the value [A] replicated to every node.:

[A], [A], [A]

If instead of removing data we add a tombstone record, our single node failure situation will look like this.:

[A, Tombstone[A]], [A, Tombstone[A]], [A]

Now when we issue a repair the Tombstone will be copied to the replica, rather than the deleted data being resurrected.:

[A, Tombstone[A]], [A, Tombstone[A]], [A, Tombstone[A]]

Our repair operation will correctly put the state of the system to what we expect with the record [A] marked as deleted on all nodes. This does mean we will end up accruing Tombstones which will permanently accumulate disk space. To avoid keeping tombstones forever we have a parameter known as gc_grace_seconds for every table in Cassandra.

The gc_grace_seconds parameter and Tombstone Removal

The table level gc_grace_seconds parameter controls how long Cassandra will retain tombstones through compaction events before finally removing them. This duration should directly reflect the amount of time a user expects to allow before recovering a failed node. After gc_grace_seconds has expired the tombstone may be removed (meaning there will no longer be any record that a certain piece of data was deleted), but as a tombstone can live in one sstable and the data it covers in another, a compaction must also include both sstable for a tombstone to be removed. More precisely, to be able to drop an actual tombstone the following needs to be true;

  • The tombstone must be older than gc_grace_seconds
  • If partition X contains the tombstone, the sstable containing the partition plus all sstables containing data older than the tombstone containing X must be included in the same compaction. We don’t need to care if the partition is in an sstable if we can guarantee that all data in that sstable is newer than the tombstone. If the tombstone is older than the data it cannot shadow that data.
  • If the option only_purge_repaired_tombstones is enabled, tombstones are only removed if the data has also been repaired.

If a node remains down or disconnected for longer than gc_grace_seconds it’s deleted data will be repaired back to the other nodes and re-appear in the cluster. This is basically the same as in the “Deletes without Tombstones” section. Note that tombstones will not be removed until a compaction event even if gc_grace_seconds has elapsed.

The default value for gc_grace_seconds is 864000 which is equivalent to 10 days. This can be set when creating or altering a table using WITH gc_grace_seconds.


Data in Cassandra can have an additional property called time to live - this is used to automatically drop data that has expired once the time is reached. Once the TTL has expired the data is converted to a tombstone which stays around for at least gc_grace_seconds. Note that if you mix data with TTL and data without TTL (or just different length of the TTL) Cassandra will have a hard time dropping the tombstones created since the partition might span many sstables and not all are compacted at once.

Fully expired sstables

If an sstable contains only tombstones and it is guaranteed that that sstable is not shadowing data in any other sstable compaction can drop that sstable. If you see sstables with only tombstones (note that TTL:ed data is considered tombstones once the time to live has expired) but it is not being dropped by compaction, it is likely that other sstables contain older data. There is a tool called sstableexpiredblockers that will list which sstables are droppable and which are blocking them from being dropped. This is especially useful for time series compaction with TimeWindowCompactionStrategy (and the deprecated DateTieredCompactionStrategy). With TimeWindowCompactionStrategy it is possible to remove the guarantee (not check for shadowing data) by enabling unsafe_aggressive_sstable_expiration.

Repaired/unrepaired data

With incremental repairs Cassandra must keep track of what data is repaired and what data is unrepaired. With anticompaction repaired data is split out into repaired and unrepaired sstables. To avoid mixing up the data again separate compaction strategy instances are run on the two sets of data, each instance only knowing about either the repaired or the unrepaired sstables. This means that if you only run incremental repair once and then never again, you might have very old data in the repaired sstables that block compaction from dropping tombstones in the unrepaired (probably newer) sstables.

Data directories

Since tombstones and data can live in different sstables it is important to realize that losing an sstable might lead to data becoming live again - the most common way of losing sstables is to have a hard drive break down. To avoid making data live tombstones and actual data are always in the same data directory. This way, if a disk is lost, all versions of a partition are lost and no data can get undeleted. To achieve this a compaction strategy instance per data directory is run in addition to the compaction strategy instances containing repaired/unrepaired data, this means that if you have 4 data directories there will be 8 compaction strategy instances running. This has a few more benefits than just avoiding data getting undeleted:

  • It is possible to run more compactions in parallel - leveled compaction will have several totally separate levelings and each one can run compactions independently from the others.
  • Users can backup and restore a single data directory.
  • Note though that currently all data directories are considered equal, so if you have a tiny disk and a big disk backing two data directories, the big one will be limited the by the small one. One work around to this is to create more data directories backed by the big disk.

Single sstable tombstone compaction

When an sstable is written a histogram with the tombstone expiry times is created and this is used to try to find sstables with very many tombstones and run single sstable compaction on that sstable in hope of being able to drop tombstones in that sstable. Before starting this it is also checked how likely it is that any tombstones will actually will be able to be dropped how much this sstable overlaps with other sstables. To avoid most of these checks the compaction option unchecked_tombstone_compaction can be enabled.

Common options

There is a number of common options for all the compaction strategies;

enabled (default: true)
Whether minor compactions should run. Note that you can have ‘enabled’: true as a compaction option and then do ‘nodetool enableautocompaction’ to start running compactions.
tombstone_threshold (default: 0.2)
How much of the sstable should be tombstones for us to consider doing a single sstable compaction of that sstable.
tombstone_compaction_interval (default: 86400s (1 day))
Since it might not be possible to drop any tombstones when doing a single sstable compaction we need to make sure that one sstable is not constantly getting recompacted - this option states how often we should try for a given sstable.
log_all (default: false)
New detailed compaction logging, see below.
unchecked_tombstone_compaction (default: false)
The single sstable compaction has quite strict checks for whether it should be started, this option disables those checks and for some usecases this might be needed. Note that this does not change anything for the actual compaction, tombstones are only dropped if it is safe to do so - it might just rewrite an sstable without being able to drop any tombstones.
only_purge_repaired_tombstone (default: false)
Option to enable the extra safety of making sure that tombstones are only dropped if the data has been repaired.
min_threshold (default: 4)
Lower limit of number of sstables before a compaction is triggered. Not used for LeveledCompactionStrategy.
max_threshold (default: 32)
Upper limit of number of sstables before a compaction is triggered. Not used for LeveledCompactionStrategy.

Further, see the section on each strategy for specific additional options.

Compaction nodetool commands

The nodetool utility provides a number of commands related to compaction:

Enable compaction.
Disable compaction.
How fast compaction should run at most - defaults to 16MB/s, but note that it is likely not possible to reach this throughput.
Statistics about current and pending compactions.
List details about the last compactions.
Set the min/max sstable count for when to trigger compaction, defaults to 4/32.

Switching the compaction strategy and options using JMX

It is possible to switch compaction strategies and its options on just a single node using JMX, this is a great way to experiment with settings without affecting the whole cluster. The mbean is:


and the attribute to change is CompactionParameters or CompactionParametersJson if you use jconsole or jmc. The syntax for the json version is the same as you would use in an ALTER TABLE statement - for example:

{ 'class': 'LeveledCompactionStrategy', 'sstable_size_in_mb': 123, 'fanout_size': 10}

The setting is kept until someone executes an ALTER TABLE that touches the compaction settings or restarts the node.

More detailed compaction logging

Enable with the compaction option log_all and a more detailed compaction log file will be produced in your log directory.

Size Tiered Compaction Strategy

The basic idea of SizeTieredCompactionStrategy (STCS) is to merge sstables of approximately the same size. All sstables are put in different buckets depending on their size. An sstable is added to the bucket if size of the sstable is within bucket_low and bucket_high of the current average size of the sstables already in the bucket. This will create several buckets and the most interesting of those buckets will be compacted. The most interesting one is decided by figuring out which bucket’s sstables takes the most reads.

Major compaction

When running a major compaction with STCS you will end up with two sstables per data directory (one for repaired data and one for unrepaired data). There is also an option (-s) to do a major compaction that splits the output into several sstables. The sizes of the sstables are approximately 50%, 25%, 12.5%… of the total size.

STCS options

min_sstable_size (default: 50MB)
Sstables smaller than this are put in the same bucket.
bucket_low (default: 0.5)
How much smaller than the average size of a bucket a sstable should be before not being included in the bucket. That is, if bucket_low * avg_bucket_size < sstable_size (and the bucket_high condition holds, see below), then the sstable is added to the bucket.
bucket_high (default: 1.5)
How much bigger than the average size of a bucket a sstable should be before not being included in the bucket. That is, if sstable_size < bucket_high * avg_bucket_size (and the bucket_low condition holds, see above), then the sstable is added to the bucket.


Defragmentation is done when many sstables are touched during a read. The result of the read is put in to the memtable so that the next read will not have to touch as many sstables. This can cause writes on a read-only-cluster.

Leveled Compaction Strategy

The idea of LeveledCompactionStrategy (LCS) is that all sstables are put into different levels where we guarantee that no overlapping sstables are in the same level. By overlapping we mean that the first/last token of a single sstable are never overlapping with other sstables. This means that for a SELECT we will only have to look for the partition key in a single sstable per level. Each level is 10x the size of the previous one and each sstable is 160MB by default. L0 is where sstables are streamed/flushed - no overlap guarantees are given here.

When picking compaction candidates we have to make sure that the compaction does not create overlap in the target level. This is done by always including all overlapping sstables in the next level. For example if we select an sstable in L3, we need to guarantee that we pick all overlapping sstables in L4 and make sure that no currently ongoing compactions will create overlap if we start that compaction. We can start many parallel compactions in a level if we guarantee that we wont create overlap. For L0 -> L1 compactions we almost always need to include all L1 sstables since most L0 sstables cover the full range. We also can’t compact all L0 sstables with all L1 sstables in a single compaction since that can use too much memory.

When deciding which level to compact LCS checks the higher levels first (with LCS, a “higher” level is one with a higher number, L0 being the lowest one) and if the level is behind a compaction will be started in that level.

Major compaction

It is possible to do a major compaction with LCS - it will currently start by filling out L1 and then once L1 is full, it continues with L2 etc. This is sub optimal and will change to create all the sstables in a high level instead, CASSANDRA-11817.


During bootstrap sstables are streamed from other nodes. The level of the remote sstable is kept to avoid many compactions after the bootstrap is done. During bootstrap the new node also takes writes while it is streaming the data from a remote node - these writes are flushed to L0 like all other writes and to avoid those sstables blocking the remote sstables from going to the correct level, we only do STCS in L0 until the bootstrap is done.

STCS in L0

If LCS gets very many L0 sstables reads are going to hit all (or most) of the L0 sstables since they are likely to be overlapping. To more quickly remedy this LCS does STCS compactions in L0 if there are more than 32 sstables there. This should improve read performance more quickly compared to letting LCS do its L0 -> L1 compactions. If you keep getting too many sstables in L0 it is likely that LCS is not the best fit for your workload and STCS could work out better.

Starved sstables

If a node ends up with a leveling where there are a few very high level sstables that are not getting compacted they might make it impossible for lower levels to drop tombstones etc. For example, if there are sstables in L6 but there is only enough data to actually get a L4 on the node the left over sstables in L6 will get starved and not compacted. This can happen if a user changes sstable_size_in_mb from 5MB to 160MB for example. To avoid this LCS tries to include those starved high level sstables in other compactions if there has been 25 compaction rounds where the highest level has not been involved.

LCS options

sstable_size_in_mb (default: 160MB)
The target compressed (if using compression) sstable size - the sstables can end up being larger if there are very large partitions on the node.
fanout_size (default: 10)
The target size of levels increases by this fanout_size multiplier. You can reduce the space amplification by tuning this option.

LCS also support the cassandra.disable_stcs_in_l0 startup option (-Dcassandra.disable_stcs_in_l0=true) to avoid doing STCS in L0.

Time Window CompactionStrategy

TimeWindowCompactionStrategy (TWCS) is designed specifically for workloads where it’s beneficial to have data on disk grouped by the timestamp of the data, a common goal when the workload is time-series in nature or when all data is written with a TTL. In an expiring/TTL workload, the contents of an entire SSTable likely expire at approximately the same time, allowing them to be dropped completely, and space reclaimed much more reliably than when using SizeTieredCompactionStrategy or LeveledCompactionStrategy. The basic concept is that TimeWindowCompactionStrategy will create 1 sstable per file for a given window, where a window is simply calculated as the combination of two primary options:

compaction_window_unit (default: DAYS)
A Java TimeUnit (MINUTES, HOURS, or DAYS).
compaction_window_size (default: 1)
The number of units that make up a window.
unsafe_aggressive_sstable_expiration (default: false)
Expired sstables will be dropped without checking its data is shadowing other sstables. This is a potentially risky option that can lead to data loss or deleted data re-appearing, going beyond what unchecked_tombstone_compaction does for single sstable compaction. Due to the risk the jvm must also be started with -Dcassandra.unsafe_aggressive_sstable_expiration=true.

Taken together, the operator can specify windows of virtually any size, and TimeWindowCompactionStrategy will work to create a single sstable for writes within that window. For efficiency during writing, the newest window will be compacted using SizeTieredCompactionStrategy.

Ideally, operators should select a compaction_window_unit and compaction_window_size pair that produces approximately 20-30 windows - if writing with a 90 day TTL, for example, a 3 Day window would be a reasonable choice ('compaction_window_unit':'DAYS','compaction_window_size':3).

TimeWindowCompactionStrategy Operational Concerns

The primary motivation for TWCS is to separate data on disk by timestamp and to allow fully expired SSTables to drop more efficiently. One potential way this optimal behavior can be subverted is if data is written to SSTables out of order, with new data and old data in the same SSTable. Out of order data can appear in two ways:

  • If the user mixes old data and new data in the traditional write path, the data will be comingled in the memtables and flushed into the same SSTable, where it will remain comingled.
  • If the user’s read requests for old data cause read repairs that pull old data into the current memtable, that data will be comingled and flushed into the same SSTable.

While TWCS tries to minimize the impact of comingled data, users should attempt to avoid this behavior. Specifically, users should avoid queries that explicitly set the timestamp via CQL USING TIMESTAMP. Additionally, users should run frequent repairs (which streams data in such a way that it does not become comingled).

Changing TimeWindowCompactionStrategy Options

Operators wishing to enable TimeWindowCompactionStrategy on existing data should consider running a major compaction first, placing all existing data into a single (old) window. Subsequent newer writes will then create typical SSTables as expected.

Operators wishing to change compaction_window_unit or compaction_window_size can do so, but may trigger additional compactions as adjacent windows are joined together. If the window size is decrease d (for example, from 24 hours to 12 hours), then the existing SSTables will not be modified - TWCS can not split existing SSTables into multiple windows.