From: Konstantin Osipov <kostja@tarantool.org>
To: tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH 02/10] ddl: synchronize user cache with actual data state
Date: Thu, 4 Jul 2019 10:42:07 +0300 [thread overview]
Message-ID: <20190704074207.GA24820@atlas> (raw)
In-Reply-To: <20190703200032.hlbegritnbzbobdn@esperanza>
* Vladimir Davydov <vdavydov.dev@gmail.com> [19/07/03 23:05]:
> > > See also commit 22bedebe715c ("ddl: synchronize privileges cache with
> > > actual data state").
> >
> > Here's a problem with replacing before commit:
> >
> > Imagine you have a yielding transaction, which is in progress.
> >
> > You replace state in the metadata cache, it doesn't matter what
> > cache - user, view, space, trigger, or whatever.
> >
> > Then you yield on DDL.
>
> If we yield from a DDL transaction, we must invoke on_rollback triggers
> immediately. I don't see any problem with that. If we don't do that now,
> I'll change that soon enough.
any transaction yields on WAL write. With synchronous replication
it will also involve a network round trip to get a quorum
responses - so a failure is more likely to happen. And then we
have a cascading rollback.
--
Konstantin Osipov, Moscow, Russia
next prev parent reply other threads:[~2019-07-04 7:42 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-03 19:30 [PATCH 00/10] Prepare box/alter.cc for transactional DDL Vladimir Davydov
2019-07-03 19:30 ` [PATCH 01/10] ddl: unreference view on space drop synchronously Vladimir Davydov
2019-07-03 19:37 ` Konstantin Osipov
2019-07-03 19:30 ` [PATCH 02/10] ddl: synchronize user cache with actual data state Vladimir Davydov
2019-07-03 19:43 ` Konstantin Osipov
2019-07-03 20:00 ` Vladimir Davydov
2019-07-04 7:42 ` Konstantin Osipov [this message]
2019-07-03 19:30 ` [PATCH 03/10] ddl: synchronize func " Vladimir Davydov
2019-07-04 8:12 ` Konstantin Osipov
2019-07-03 19:30 ` [PATCH 04/10] ddl: synchronize sequence " Vladimir Davydov
2019-07-04 8:16 ` Konstantin Osipov
2019-07-03 19:30 ` [PATCH 05/10] ddl: fix _space_sequence rollback Vladimir Davydov
2019-07-03 19:30 ` [PATCH 06/10] ddl: restore sequence value if drop is rolled back Vladimir Davydov
2019-07-03 19:30 ` [PATCH 07/10] ddl: don't use txn_last_stmt on _collation commit/rollback Vladimir Davydov
2019-07-03 19:30 ` [PATCH 08/10] ddl: don't use txn_last_stmt on _trigger commit/rollback Vladimir Davydov
2019-07-03 19:30 ` [PATCH 09/10] ddl: don't use txn_last_stmt on _ck_constraint commit/rollback Vladimir Davydov
2019-07-03 19:30 ` [PATCH 10/10] ddl: don't use txn_last_stmt on _cluster commit/rollback Vladimir Davydov
2019-07-04 15:01 ` [PATCH 00/10] Prepare box/alter.cc for transactional DDL 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=20190704074207.GA24820@atlas \
--to=kostja@tarantool.org \
--cc=tarantool-patches@freelists.org \
--subject='[tarantool-patches] Re: [PATCH 02/10] ddl: synchronize user cache with actual data state' \
/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