Skip to:

  1. Skip to Jira Navigation
  2. Skip to Side Navigation
  3. Skip to Main Content

GSI parallel compaction

Description

Currently, GSI is limited to performing compaction on one index per node at a time. In the case that there are multiple indexes on one node, this can lead to disk use constantly growing as by the time one index has been compacted, the others have grown by a greater amount than was compacted. This is the case even when index traffic is within GSI's limitations.

This should be less of an issue with circular compaction in later versions, but still may be a useful feature as full compaction will be required from time to time.

An important caveat to note is the impact on diskIO of multiple compactors, see .

Components

Affects versions

Fix versions

Labels

Environment

None

Release Notes Description

None

Activity

Matt Ingenthron July 26, 2017 at 8:51 PM

Reassigning these to Chin to be given to the right owner.

John Liang March 21, 2016 at 6:31 PM
Edited

We need performance data from forestdb on how much improvement we will get from parallelized compaction. With parallel compaction, it will use up more I/O on compaction and it is not clear how this affect concurrent write/read/commit traffic.

Sriram Melkote March 21, 2016 at 6:28 PM

Also, we've not established this will actually help the customer. So it's not yet committed to fix in watson or elsewhere yet.

Wayne Siu March 21, 2016 at 6:27 PM

As discussed in Maintenance meeting, "improvement" ticket will be assigned to PM to review.

Pinned fields
Click on the next to a field label to start pinning.

Details

Assignee

Reporter

Priority

Instabug

Open Instabug

PagerDuty

Sentry

Zendesk Support

Created March 16, 2016 at 10:21 AM
Updated April 13, 2018 at 4:54 AM
Instabug