Tarantool development patches archive
 help / color / mirror / Atom feed
From: Cyrill Gorcunov via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: Konstantin Osipov <kostja.osipov@gmail.com>
Cc: Mons Anderson <v.perepelitsa@corp.mail.ru>,
	tml <tarantool-patches@dev.tarantool.org>,
	Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
Subject: Re: [Tarantool-patches] [RFC] gc/xlog: delay xlog cleanup until relays are subscribed
Date: Thu, 18 Mar 2021 10:41:21 +0300	[thread overview]
Message-ID: <YFMEIbLBC/rrNbXK@grain> (raw)
In-Reply-To: <20210318051812.GB224985@starling>

On Thu, Mar 18, 2021 at 08:18:12AM +0300, Konstantin Osipov wrote:
> > > 
> > > Should be in 1.10 as well.
> > 
> > Sure, but first I need to be sure if current rfc is acceptable
> > in general and I didn't miss something. I suspect we might
> > need to extend this code and better to not make some design
> > mistakes which gonna be hard to resolve later. I've a test
> > case for this simply didn't posted it yet.
> 
> This is a case where a configuration setting would be suitable. I
> don't think it's necessary to keep these logs forever by default,
> 4-6 hours should be a good default for many setups. Some will set
> the default to, say, 5 minutes, essentially to keep the old
> behaviour, and some can set it to infinity, to get the current
> behaviour of the patch.
> 
> All tarantool setting follow the convention to start from
> subsystem name, so it got to be wal_{something}, e.g.
> wal_keep_logs

Good point, thanks! Kostya, lets clarify some moments:

1) We introduce "wal_keep_logs" option which defines
   a timeout to kick the cleanup fiber.
2) If node is anonymous replica we simply ignore this
   option.
3) If this option is set then we have a few subcases:
   a) The `_cluster` space is not empty thus thus once
      all replicas are subscribed _before_ the timeout
      expired we trigger the cleanup fiber since it is
      safe to process;
   b) If replicas are not connected and timeout is
      expired we kick the cleanup fiber;

Or you mean to always conside "wal_keep_logs" option
and never trigger the cleanup until it get expired?

  reply	other threads:[~2021-03-18  7:41 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-17 18:57 Cyrill Gorcunov via Tarantool-patches
2021-03-17 18:59 ` Cyrill Gorcunov via Tarantool-patches
2021-03-17 21:23 ` Konstantin Osipov via Tarantool-patches
2021-03-17 21:53   ` Cyrill Gorcunov via Tarantool-patches
2021-03-18  5:18     ` Konstantin Osipov via Tarantool-patches
2021-03-18  7:41       ` Cyrill Gorcunov via Tarantool-patches [this message]
2021-03-18  7:51         ` Konstantin Osipov via Tarantool-patches
2021-03-18  7:56           ` Cyrill Gorcunov via Tarantool-patches
2021-03-18 20:36           ` Vladislav Shpilevoy via Tarantool-patches
2021-03-18 20:45             ` Cyrill Gorcunov via Tarantool-patches
2021-03-18 20:54               ` Vladislav Shpilevoy via Tarantool-patches
2021-03-18 21:31                 ` Cyrill Gorcunov via Tarantool-patches
2021-03-18 23:04                   ` Vladislav Shpilevoy via Tarantool-patches
2021-03-19  7:13                     ` Cyrill Gorcunov via Tarantool-patches
2021-03-19  8:00                     ` Konstantin Osipov via Tarantool-patches
2021-03-19  7:40             ` Konstantin Osipov via Tarantool-patches

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=YFMEIbLBC/rrNbXK@grain \
    --to=tarantool-patches@dev.tarantool.org \
    --cc=gorcunov@gmail.com \
    --cc=kostja.osipov@gmail.com \
    --cc=v.perepelitsa@corp.mail.ru \
    --cc=v.shpilevoy@tarantool.org \
    --subject='Re: [Tarantool-patches] [RFC] gc/xlog: delay xlog cleanup until relays are subscribed' \
    /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