Tarantool development patches archive
 help / color / mirror / Atom feed
From: Vladimir Davydov <vdavydov.dev@gmail.com>
To: Georgy Kirichenko <georgy@tarantool.org>
Cc: tarantool-patches@freelists.org
Subject: Re: [tarantool-patches] [PATCH v2 3/5] Enforce applier out of order protection
Date: Mon, 28 Jan 2019 15:09:01 +0300	[thread overview]
Message-ID: <20190128120901.spkitg7kyrfjp6xz@esperanza> (raw)
In-Reply-To: <4c39bbbfcd12c47b9b14fc1a0a0484331939ed63.1548152776.git.georgy@tarantool.org>

On Tue, Jan 22, 2019 at 01:31:11PM +0300, Georgy Kirichenko wrote:
> Do not skip row until the row is not processed by other appliers.

Looks like a fix for

  https://github.com/tarantool/tarantool/issues/3568

Worth adding a test?

> 
> Prerequisite #980
> ---
>  src/box/applier.cc | 35 ++++++++++++++++++-----------------
>  1 file changed, 18 insertions(+), 17 deletions(-)
> 
> diff --git a/src/box/applier.cc b/src/box/applier.cc
> index 87873e970..148c8ce5a 100644
> --- a/src/box/applier.cc
> +++ b/src/box/applier.cc
> @@ -504,6 +504,22 @@ applier_subscribe(struct applier *applier)
>  
>  		applier->lag = ev_now(loop()) - row.tm;
>  		applier->last_row_time = ev_monotonic_now(loop());
> +		struct replica *replica = replica_by_id(row.replica_id);
> +		struct latch *latch = (replica ? &replica->order_latch :
> +				       &replicaset.applier.order_latch);
> +		/*
> +		 * In a full mesh topology, the same set
> +		 * of changes may arrive via two
> +		 * concurrently running appliers. Thanks
> +		 * to vclock_follow() above, the first row

I don't see any vclock_follow() above. Please fix the comment.

> +		 * in the set will be skipped - but the
> +		 * remaining may execute out of order,
> +		 * when the following xstream_write()
> +		 * yields on WAL. Hence we need a latch to
> +		 * strictly order all changes which belong
> +		 * to the same server id.
> +		 */
> +		latch_lock(latch);
>  		if (vclock_get(&replicaset.applier.vclock,
>  			       row.replica_id) < row.lsn) {
>  			if (row.replica_id == instance_id &&

AFAIU this patch makes replicaset.applier.vclock, introduced by the
previous patch, useless.

> @@ -516,24 +532,7 @@ applier_subscribe(struct applier *applier)
>  			int64_t old_lsn = vclock_get(&replicaset.applier.vclock,
>  						     row.replica_id);
>  			vclock_follow_xrow(&replicaset.applier.vclock, &row);
> -			struct replica *replica = replica_by_id(row.replica_id);
> -			struct latch *latch = (replica ? &replica->order_latch :
> -					       &replicaset.applier.order_latch);
> -			/*
> -			 * In a full mesh topology, the same set
> -			 * of changes may arrive via two
> -			 * concurrently running appliers. Thanks
> -			 * to vclock_follow() above, the first row
> -			 * in the set will be skipped - but the
> -			 * remaining may execute out of order,
> -			 * when the following xstream_write()
> -			 * yields on WAL. Hence we need a latch to
> -			 * strictly order all changes which belong
> -			 * to the same server id.
> -			 */
> -			latch_lock(latch);
>  			int res = xstream_write(applier->subscribe_stream, &row);
> -			latch_unlock(latch);
>  			if (res != 0) {
>  				struct error *e = diag_last_error(diag_get());
>  				/**
> @@ -548,11 +547,13 @@ applier_subscribe(struct applier *applier)
>  					/* Rollback lsn to have a chance for a retry. */
>  					vclock_set(&replicaset.applier.vclock,
>  						   row.replica_id, old_lsn);
> +					latch_unlock(latch);
>  					diag_raise();
>  				}
>  			}
>  		}
>  done:
> +		latch_unlock(latch);
>  		/*
>  		 * Stay 'orphan' until appliers catch up with
>  		 * the remote vclock at the time of SUBSCRIBE

  reply	other threads:[~2019-01-28 12:09 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-01-22 10:31 [tarantool-patches] [PATCH v2 0/5] Strong sequentially LSN in journal Georgy Kirichenko
2019-01-22 10:31 ` [tarantool-patches] [PATCH v2 1/5] Do not promote wal vclock for failed writes Georgy Kirichenko
2019-01-28 11:20   ` Vladimir Davydov
2019-01-29 10:22     ` Георгий Кириченко
2019-01-29 11:58       ` Vladimir Davydov
2019-01-22 10:31 ` [tarantool-patches] [PATCH v2 2/5] Update replicaset vclock from wal Georgy Kirichenko
2019-01-28 11:59   ` Vladimir Davydov
2019-01-29 10:33     ` [tarantool-patches] " Георгий Кириченко
2019-01-22 10:31 ` [tarantool-patches] [PATCH v2 3/5] Enforce applier out of order protection Georgy Kirichenko
2019-01-28 12:09   ` Vladimir Davydov [this message]
2019-01-29 10:30     ` [tarantool-patches] " Георгий Кириченко
2019-01-29 12:00       ` Vladimir Davydov
2019-01-22 10:31 ` [tarantool-patches] [PATCH v2 4/5] Emit NOP if an applier skips row Georgy Kirichenko
2019-01-28 12:15   ` Vladimir Davydov
2019-02-08 16:50   ` [tarantool-patches] " Konstantin Osipov
2019-01-22 10:31 ` [tarantool-patches] [PATCH v2 5/5] Disallow lsn gaps while vclock following Georgy Kirichenko
2019-01-28 12:18   ` Vladimir Davydov
2019-01-28 11:15 ` [tarantool-patches] [PATCH v2 0/5] Strong sequentially LSN in journal 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=20190128120901.spkitg7kyrfjp6xz@esperanza \
    --to=vdavydov.dev@gmail.com \
    --cc=georgy@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='Re: [tarantool-patches] [PATCH v2 3/5] Enforce applier out of order protection' \
    /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