Tarantool development patches archive
 help / color / mirror / Atom feed
From: Serge Petrenko via Tarantool-patches <tarantool-patches@dev.tarantool.org>
To: Yan Shtunder <ya.shtunder@gmail.com>
Cc: tarantool-patches@dev.tarantool.org
Subject: Re: [Tarantool-patches] [PATCH] replication: the truncate method called from within a transaction
Date: Wed, 25 Aug 2021 15:45:25 +0300	[thread overview]
Message-ID: <0f8586e4-d5df-f720-de75-4af9f3fb0a3b@tarantool.org> (raw)
In-Reply-To: <20210825055451.115351-1-ya.shtunder@gmail.com>



25.08.2021 08:54, Yan Shtunder пишет:
> The truncate method could be called from within a transaction. The flag
> of GROUP_LOCAL was set in truncate method after statement row had been
> being checked on the GROUP_LOCAL. Accordingly, after a local transaction
> NOP row was not appended.
>
> Closes #6123
> ---

Hi! Thanks for the patch!

Please, find a couple of comments below.

> Issue: https://github.com/tarantool/tarantool/issues/6123
> Patch: https://github.com/tarantool/tarantool/tree/yshtunder/gh-6123-truncate-is-local-transaction
>
>   src/box/txn.c                                 | 16 +++--
>   ...-6123-truncate-is-local-transaction.result | 69 +++++++++++++++++++
>   ...123-truncate-is-local-transaction.test.lua | 27 ++++++++
>   3 files changed, 107 insertions(+), 5 deletions(-)
>   create mode 100644 test/replication/gh-6123-truncate-is-local-transaction.result
>   create mode 100644 test/replication/gh-6123-truncate-is-local-transaction.test.lua
>
> diff --git a/src/box/txn.c b/src/box/txn.c
> index b80e722a4..d37bec7ae 100644
> --- a/src/box/txn.c
> +++ b/src/box/txn.c
> @@ -397,6 +397,7 @@ txn_is_distributed(struct txn *txn)
>   int
>   txn_commit_stmt(struct txn *txn, struct request *request)
>   {
> +	int rc = 0;
>   	assert(txn->in_sub_stmt > 0);
>   	/*
>   	 * Run on_replace triggers. For now, disallow mutation
> @@ -415,9 +416,6 @@ txn_commit_stmt(struct txn *txn, struct request *request)
>   		assert(stmt->row != NULL);
>   		if (stmt->row->replica_id == 0) {
>   			++txn->n_new_rows;
> -			if (stmt->row->group_id == GROUP_LOCAL)
> -				++txn->n_local_rows;
> -
>   		} else {
>   			++txn->n_applier_rows;
>   		}

Why can't you move this whole block lower, not just the if() body?

I mean both if() {} and else {}
> @@ -447,10 +445,18 @@ txn_commit_stmt(struct txn *txn, struct request *request)
>   			 */
>   			stmt->does_require_old_tuple = true;
>
> -			if(trigger_run(&stmt->space->on_replace, txn) != 0)
> -				goto fail;
> +			rc = trigger_run(&stmt->space->on_replace, txn);
>   		}
>   	}
> +
> +	if (stmt->row != NULL && stmt->row->replica_id == 0 &&
> +	    stmt->row->group_id == GROUP_LOCAL) {
> +		++txn->n_local_rows;
> +	}
> +
> +	if (rc != 0)
> +		goto fail;
> +
>   	--txn->in_sub_stmt;
>   	return 0;

This could be written simpler:

if (rc != 0) {
      // the code below the "fail" label goes here
} else {
     --txn->in_sub_stmt;
}
return rc;
>   fail:
> diff --git a/test/replication/gh-6123-truncate-is-local-transaction.result b/test/replication/gh-6123-truncate-is-local-transaction.result
> new file mode 100644
> index 000000000..0a4b990c1
> --- /dev/null
> +++ b/test/replication/gh-6123-truncate-is-local-transaction.result
> @@ -0,0 +1,69 @@
> +-- test-run result file version 2
> +test_run = require('test_run').new()
> + | ---
> + | ...
> +
> +
> +-- Step 1
> +box.schema.user.grant("guest", "replication")
> + | ---
> + | ...
> +s = box.schema.space.create("temp", {temporary=true})
> + | ---
> + | ...
> +_ = s:create_index('pk')
> + | ---
> + | ...
> +s:insert{1,2}
> + | ---
> + | - [1, 2]
> + | ...
> +s:insert{4}
> + | ---
> + | - [4]
> + | ...
> +s:select()
> + | ---
> + | - - [1, 2]
> + |   - [4]
> + | ...
> +
> +
> +-- Step 2
> +test_run:cmd('create server replica with rpl_master=default,\
> +              script="replication/replica.lua"')
> + | ---
> + | - true
> + | ...
> +test_run:cmd('start server replica')
> + | ---
> + | - true
> + | ...
> +
> +
> +-- Step 3
> +box.begin() box.space._schema:replace{"smth"} s:truncate() box.commit()
> + | ---
> + | ...
> +s:select()
> + | ---
> + | - []
> + | ...
> +box.space._schema:select{'smth'}
> + | ---
> + | - - ['smth']
> + | ...
> +
> +
> +test_run:cmd('stop server replica')
> + | ---
> + | - true
> + | ...
> +test_run:cmd('cleanup server replica')
> + | ---
> + | - true
> + | ...
> +test_run:cmd('delete server replica')
> + | ---
> + | - true
> + | ...
> diff --git a/test/replication/gh-6123-truncate-is-local-transaction.test.lua b/test/replication/gh-6123-truncate-is-local-transaction.test.lua
> new file mode 100644
> index 000000000..b68514d0b
> --- /dev/null
> +++ b/test/replication/gh-6123-truncate-is-local-transaction.test.lua
> @@ -0,0 +1,27 @@
> +test_run = require('test_run').new()
> +
> +
> +-- Step 1
> +box.schema.user.grant("guest", "replication")
> +s = box.schema.space.create("temp", {temporary=true})
> +_ = s:create_index('pk')
> +s:insert{1,2}
> +s:insert{4}
> +s:select()
> +
> +
> +-- Step 2
> +test_run:cmd('create server replica with rpl_master=default,\
> +              script="replication/replica.lua"')
> +test_run:cmd('start server replica')
> +
> +
> +-- Step 3
> +box.begin() box.space._schema:replace{"smth"} s:truncate() box.commit()
> +s:select()
> +box.space._schema:select{'smth'}
> +
> +
> +test_run:cmd('stop server replica')
> +test_run:cmd('cleanup server replica')
> +test_run:cmd('delete server replica')
> --
> 2.25.1
>
The test looks good with one comment:

Before stopping the replica, you should check that replica has
received the last transaction, and that replication isn't broken.

-- 
Serge Petrenko


  reply	other threads:[~2021-08-25 12:45 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-25  5:54 Yan Shtunder via Tarantool-patches
2021-08-25 12:45 ` Serge Petrenko via Tarantool-patches [this message]
2021-08-27 12:19 Yan Shtunder via Tarantool-patches
2021-09-21 16:31 Yan Shtunder via Tarantool-patches
2021-09-27  6:31 ` Serge Petrenko 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=0f8586e4-d5df-f720-de75-4af9f3fb0a3b@tarantool.org \
    --to=tarantool-patches@dev.tarantool.org \
    --cc=sergepetrenko@tarantool.org \
    --cc=ya.shtunder@gmail.com \
    --subject='Re: [Tarantool-patches] [PATCH] replication: the truncate method called from within a transaction' \
    /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