From: Kristian Nielsen
Date: July 6 2011 5:51pm
Subject: Re: "OS thread id" removed
List-Archive: http://lists.mysql.com/internals/38354
Message-Id: <87oc17wb5l.fsf@knielsen-hq.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

ghazel@stripped writes:

> The "OS thread id" section of TRANSACTION output from "SHOW ENGINE
> INNODB STATUS" appears to have been removed. There was no changelog
> entry for this that I could find, and Launchpad only has this
> unhelpful "merge from trunk" commit message with a huge pile of other
> unrelated changes:
> http://bazaar.launchpad.net/~mysql/mysql-server/mysql-trunk/revision/3071
>
> First off, where can I find a real commit message?

    bzr branch lp:mysql-server
    bzr log -p --show-ids -rrevid:marko.makela@stripped=
6mbq0wokqdxl9m

revno: 3356
revision-id: marko.makela@stripped
parent: vasil.dimov@stripped
committer: Marko M=C3=A4kel=C3=A4 <marko.makela@stripped>
branch nick: 5.5-innodb
timestamp: Tue 2011-04-05 10:18:43 +0300
message:
  Bug 12323643 - CLEAN UP THE INNODB THREAD SHUTDOWN AND ASSERTIONS (WL#513=
6)
=20=20
  Remove most references to thread id in InnoDB. Three references
  remain: the current holder of a mutex, and the current x-lock holder
  of a rw-lock, and some references in UNIV_SYNC_DEBUG checks. This
  allows MySQL to change the thread associated to a client connection.
=20=20
  Tighten the UNIV_SYNC_DEBUG checks, trying to ensure that no InnoDB
  mutex or x-lock is being held when returning control to MySQL. The
  only semaphore that may be held is the btr_search_latch in shared mode.
=20=20
  sync_thread_levels_empty_except_dict(): A wrapper for
  sync_thread_levels_empty_gen(TRUE).
=20=20
  sync_thread_levels_nonempty_trx(): Check that the current thread is
  not holding any InnoDB semaphores, except btr_search_latch if
  trx->has_search_latch.
=20=20
  sync_thread_levels_empty(): Unused function; remove.
=20=20
  trx_t: Remove mysql_thread_id and mysql_process_no.
=20=20
  srv_slot_t: Remove id and handle.
=20=20
  row_search_for_mysql(), srv_conc_enter_innodb(),
  srv_conc_force_enter_innodb(), srv_conc_force_exit_innodb(),
  srv_conc_exit_innodb(), srv_suspend_mysql_thread: Assert
  !sync_thread_levels_nonempty_trx().
=20=20
  rb:634 approved by Sunny Bains

I do not know if there is a way to get this directly from Launchpad, probab=
ly not.

> Second, is there no
> way to get the OS thread id now? It was very handy when tying MySQL
> thread id to OS thread id for debugging with gdb.

I don't know of any easy way, it looks like the thread id was removed from =
the
data structures.

Note that the bug referenced is in the new internal bug tracker, which is n=
ot
accessible. And the WL#5136 is marked private. Welcome to the new style of
Open Source development in MySQL ... but maybe an Oracle developer will
comment, some of them are quite helpful.

If I should guess, this is in preparation of better support for thread pool=
. I
heard rumours of a (closed-source) thread pool being developed at Oracle. W=
ith
a good thread pool, different statements in the same transaction can run in
different OS threads (or potentially even different parts of a single
statement), so I guess it no longer makes as much sense to associate a sing=
le
OS thread ID with a transaction.

Hope this helps,

 - Kristian.