From: Konstantin Osipov <kostja@tarantool.org>
To: tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH v2 6/8] vinyl: randomize range compaction to avoid IO load spikes
Date: Mon, 11 Feb 2019 21:21:57 +0300 [thread overview]
Message-ID: <20190211182157.GB10433@chai> (raw)
In-Reply-To: <1d11af88e61d4ddb86f3406202ad178c1c7672cf.1548349067.git.vdavydov.dev@gmail.com>
* Vladimir Davydov <vdavydov.dev@gmail.com> [19/01/24 20:16]:
> Since all ranges constituting an LSM tree have the same configuration,
> they tend to get compacted at approximately the same time. This entails
> IO load spikes, which, in turn, lead to deviation of the LSM tree from
> the target shape and hence increased read amplification. To prevent this
> from happening, this patch implements compaction randomization: with 10%
> probability we defer compaction at each LSM tree level, i.e. if the
> number of runs at a level exceeds the configured run_count_per_level,
> the level will be compacted with 90%-probability, but with 10%
> probability it won't - compaction will be deferred until another run
> is added to the level.
>
> Our simulations show that such a simple algorithm performs fairly well:
> it randomizes compaction pace among ranges, spreading IO load evenly in
> time, while the write amplification is increased by not more than 5-10%,
> which seems to be a reasonable price for elimination of IO load spikes.
>
> Closes #3944
OK to push.
--
Konstantin Osipov, Moscow, Russia, +7 903 626 22 32
http://tarantool.io - www.twitter.com/kostja_osipov
next prev parent reply other threads:[~2019-02-11 18:21 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-24 17:12 [PATCH v2 0/8] vinyl: compaction randomization and throttling Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 1/8] vinyl: use uncompressed run size for range split/coalesce/compaction Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 2/8] vinyl: fix compaction priority calculation Vladimir Davydov
2019-02-08 17:18 ` [tarantool-patches] " Konstantin Osipov
2019-02-11 14:42 ` Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 3/8] vinyl: rename lsm->range_heap to max_compaction_priority Vladimir Davydov
2019-02-08 17:19 ` [tarantool-patches] " Konstantin Osipov
2019-02-08 17:26 ` Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 4/8] vinyl: keep track of dumps per compaction for each LSM tree Vladimir Davydov
2019-02-08 17:42 ` Vladimir Davydov
2019-02-11 18:17 ` [tarantool-patches] " Konstantin Osipov
2019-02-11 18:17 ` Konstantin Osipov
2019-02-12 10:15 ` Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 5/8] vinyl: set range size automatically Vladimir Davydov
2019-02-11 18:21 ` [tarantool-patches] " Konstantin Osipov
2019-02-12 10:16 ` Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 6/8] vinyl: randomize range compaction to avoid IO load spikes Vladimir Davydov
2019-02-11 18:21 ` Konstantin Osipov [this message]
2019-02-12 10:16 ` [tarantool-patches] " Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 7/8] vinyl: introduce quota consumer types Vladimir Davydov
2019-02-12 15:48 ` Vladimir Davydov
2019-01-24 17:12 ` [PATCH v2 8/8] vinyl: throttle tx to ensure compaction keeps up with dumps Vladimir Davydov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190211182157.GB10433@chai \
--to=kostja@tarantool.org \
--cc=tarantool-patches@freelists.org \
--subject='[tarantool-patches] Re: [PATCH v2 6/8] vinyl: randomize range compaction to avoid IO load spikes' \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox