Tarantool development patches archive
 help / color / mirror / Atom feed
From: "Serge Petrenko" <sergepetrenko@tarantool.org>
To: "Cyrill Gorcunov" <gorcunov@gmail.com>
Cc: tml <tarantool-patches@dev.tarantool.org>,
	"Vladislav Shpilevoy" <v.shpilevoy@tarantool.org>
Subject: Re: [Tarantool-patches] [PATCH v9 4/7] qsync: direct write of CONFIRM/ROLLBACK into a journal
Date: Fri, 21 Aug 2020 11:36:57 +0300	[thread overview]
Message-ID: <1597999017.380175980@f519.i.mail.ru> (raw)
In-Reply-To: <20200819213442.1099018-5-gorcunov@gmail.com>

[-- Attachment #1: Type: text/plain, Size: 3968 bytes --]


Thanks for the patch!
LGTM.
 
  
>Четверг, 20 августа 2020, 0:35 +03:00 от Cyrill Gorcunov <gorcunov@gmail.com>:
> 
>When we need to write CONFIRM or ROLLBACK message (which is
>a binary record in msgpack format) into a journal we use txn code
>to allocate a new transaction, encode there a message and pass it
>to walk the long txn path before it hit the journal. This is not
>only resource wasting but also somehow strange from architectural
>point of view.
>
>Instead lets encode a record on the stack and write it to the journal
>directly.
>
>Part-of #5129
>
>Signed-off-by: Cyrill Gorcunov < gorcunov@gmail.com >
>---
> src/box/txn_limbo.c | 66 +++++++++++++++++++++++----------------------
> 1 file changed, 34 insertions(+), 32 deletions(-)
>
>diff --git a/src/box/txn_limbo.c b/src/box/txn_limbo.c
>index e458dad75..4b90d7fa5 100644
>--- a/src/box/txn_limbo.c
>+++ b/src/box/txn_limbo.c
>@@ -32,6 +32,7 @@
> #include "txn_limbo.h"
> #include "replication.h"
> #include "iproto_constants.h"
>+#include "journal.h"
> 
> struct txn_limbo txn_limbo;
> 
>@@ -272,6 +273,17 @@ txn_limbo_wait_complete(struct txn_limbo *limbo, struct txn_limbo_entry *entry)
>  return 0;
> }
> 
>+/**
>+ * A callback for synchronous write: txn_limbo_write_synchro fiber
>+ * waiting to proceed once a record is written to WAL.
>+ */
>+static void
>+txn_limbo_write_cb(struct journal_entry *entry)
>+{
>+ assert(entry->complete_data != NULL);
>+ fiber_wakeup(entry->complete_data);
>+}
>+
> static void
> txn_limbo_write_synchro(struct txn_limbo *limbo, uint32_t type, int64_t lsn)
> {
>@@ -285,46 +297,36 @@ txn_limbo_write_synchro(struct txn_limbo *limbo, uint32_t type, int64_t lsn)
> 
>  /*
>  * This is a synchronous commit so we can
>- * use body and row allocated on a stack.
>+ * allocate everything on a stack.
>  */
>  struct synchro_body_bin body;
>  struct xrow_header row;
>- struct request request = {
>- .header = &row,
>- };
>+ char buf[sizeof(struct journal_entry) +
>+ sizeof(struct xrow_header *)];
> 
>- struct txn *txn = txn_begin();
>- if (txn == NULL)
>- goto rollback;
>+ struct journal_entry *entry = (struct journal_entry *)buf;
>+ entry->rows[0] = &row;
> 
>  xrow_encode_synchro(&row, &body, &req);
> 
>- /*
>- * This is not really a transaction. It just uses txn API
>- * to put the data into WAL. And obviously it should not
>- * go to the limbo and block on the very same sync
>- * transaction which it tries to confirm now.
>- */
>- txn_set_flag(txn, TXN_FORCE_ASYNC);
>-
>- if (txn_begin_stmt(txn, NULL) != 0)
>- goto rollback;
>- if (txn_commit_stmt(txn, &request) != 0)
>- goto rollback;
>- if (txn_commit(txn) != 0)
>- goto rollback;
>- return;
>+ journal_entry_create(entry, 1, xrow_approx_len(&row),
>+ txn_limbo_write_cb, fiber());
> 
>-rollback:
>- /*
>- * XXX: the stub is supposed to be removed once it is defined what to do
>- * when a synchro request WAL write fails. One of the possible
>- * solutions: log the error, keep the limbo queue as is and probably put
>- * in rollback mode. Then provide a hook to call manually when WAL
>- * problems are fixed. Or retry automatically with some period.
>- */
>- panic("Could not write a synchro request to WAL: lsn = %lld, type = "
>- "%s\n", lsn, iproto_type_name(type));
>+ if (journal_write(entry) != 0 || entry->res < 0) {
>+ diag_set(ClientError, ER_WAL_IO);
>+ diag_log();
>+ /*
>+ * XXX: the stub is supposed to be removed once it is defined
>+ * what to do when a synchro request WAL write fails. One of
>+ * the possible solutions: log the error, keep the limbo
>+ * queue as is and probably put in rollback mode. Then
>+ * provide a hook to call manually when WAL problems are fixed.
>+ * Or retry automatically with some period.
>+ */
>+ panic("Could not write a synchro request to WAL: "
>+ "lsn = %lld, type = %s\n", lsn,
>+ type == IPROTO_CONFIRM ? "CONFIRM" : "ROLLBACK");
>+ }
> }
> 
> /**
>--
>2.26.2
 
 
--
Serge Petrenko  
 

[-- Attachment #2: Type: text/html, Size: 4867 bytes --]

  parent reply	other threads:[~2020-08-21  8:36 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-19 21:34 [Tarantool-patches] [PATCH v9 0/7] qsync: write CONFIRM/ROLLBACK without txn engine Cyrill Gorcunov
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 1/7] journal: bind asynchronous write completion to an entry Cyrill Gorcunov
2020-08-21  7:48   ` Serge Petrenko
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 2/7] journal: add journal_entry_create helper Cyrill Gorcunov
2020-08-21  7:51   ` Serge Petrenko
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 3/7] qsync: provide a binary form of syncro entries Cyrill Gorcunov
2020-08-21  8:15   ` Serge Petrenko
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 4/7] qsync: direct write of CONFIRM/ROLLBACK into a journal Cyrill Gorcunov
2020-08-19 22:43   ` Vladislav Shpilevoy
2020-08-20  7:13     ` Cyrill Gorcunov
2020-08-21  8:36   ` Serge Petrenko [this message]
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 5/7] applier: process synchro requests without txn engine Cyrill Gorcunov
2020-08-21  8:51   ` Serge Petrenko
2020-08-21 21:59     ` Vladislav Shpilevoy
2020-08-23 12:15       ` Serge Petrenko
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 6/7] txn: txn_add_redo -- drop synchro processing Cyrill Gorcunov
2020-08-21  8:52   ` Serge Petrenko
2020-08-19 21:34 ` [Tarantool-patches] [PATCH v9 7/7] xrow: drop xrow_header_dup_body Cyrill Gorcunov
2020-08-21  8:57   ` Serge Petrenko
2020-08-24 21:16 ` [Tarantool-patches] [PATCH v9 0/7] qsync: write CONFIRM/ROLLBACK without txn engine Vladislav Shpilevoy

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=1597999017.380175980@f519.i.mail.ru \
    --to=sergepetrenko@tarantool.org \
    --cc=gorcunov@gmail.com \
    --cc=tarantool-patches@dev.tarantool.org \
    --cc=v.shpilevoy@tarantool.org \
    --subject='Re: [Tarantool-patches] [PATCH v9 4/7] qsync: direct write of CONFIRM/ROLLBACK into a journal' \
    /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