[Maria-developers] Patches to consider to include
Hello, What do you think about including patches global_trx_ids, binlog event checksums and fixes to crash-proof slaves from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids into tree ? -- Vadim Tkachenko, CTO Percona Inc. ICQ: 369-510-335, Skype: vadimtk153, Phone +1-888-401-3403 MySQL Performance Blog - http://www.mysqlperformanceblog.com MySQL Consulting http://www.percona.com/
Generally I think it would be great if Google could use MariaDB instead of there own customized version of MySQL at some stage. At least as soon as they upgrade to 5.1. On May 14, 2009, at 7:26 AM, Vadim Tkachenko wrote:
Hello,
What do you think about including patches global_trx_ids, binlog event checksums and fixes to crash-proof slaves from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids into tree ?
-- Vadim Tkachenko, CTO Percona Inc. ICQ: 369-510-335, Skype: vadimtk153, Phone +1-888-401-3403 MySQL Performance Blog - http://www.mysqlperformanceblog.com MySQL Consulting http://www.percona.com/
_______________________________________________ Mailing list: https://launchpad.net/~maria-developers Post to : maria-developers@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-developers More help : https://help.launchpad.net/ListHelp
-- Paul McCullagh PrimeBase Technologies www.primebase.org www.blobstreaming.org pbxt.blogspot.com
Paul, I do not think it happens soon. On conference Mark said - No more upgrades :) Paul McCullagh wrote:
Generally I think it would be great if Google could use MariaDB instead of there own customized version of MySQL at some stage.
At least as soon as they upgrade to 5.1.
On May 14, 2009, at 7:26 AM, Vadim Tkachenko wrote:
Hello,
What do you think about including patches global_trx_ids, binlog event checksums and fixes to crash-proof slaves from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids into tree ?
-- Vadim Tkachenko, CTO Percona Inc. ICQ: 369-510-335, Skype: vadimtk153, Phone +1-888-401-3403 MySQL Performance Blog - http://www.mysqlperformanceblog.com MySQL Consulting http://www.percona.com/
_______________________________________________ Mailing list: https://launchpad.net/~maria-developers Post to : maria-developers@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-developers More help : https://help.launchpad.net/ListHelp
-- Paul McCullagh PrimeBase Technologies www.primebase.org www.blobstreaming.org pbxt.blogspot.com
-- Vadim Tkachenko, CTO Percona Inc. ICQ: 369-510-335, Skype: vadimtk153, Phone +1-888-401-3403 MySQL Performance Blog - http://www.mysqlperformanceblog.com MySQL Consulting http://www.percona.com/
On May 14, 2009, at 5:15 PM, Vadim Tkachenko wrote:
Paul,
I do not think it happens soon.
Probably not :(
On conference Mark said - No more upgrades :)
Yup, they are happy that their system is running and that they have all the features they require now. -- Paul McCullagh PrimeBase Technologies www.primebase.org www.blobstreaming.org pbxt.blogspot.com
On Thu, May 14, 2009 at 8:58 AM, Paul McCullagh <paul.mccullagh@primebase.org> wrote:
On May 14, 2009, at 5:15 PM, Vadim Tkachenko wrote:
Paul,
I do not think it happens soon.
Probably not :(
On conference Mark said - No more upgrades :)
Yup, they are happy that their system is running and that they have all the features they require now.
I think it will be great when there won't have to be another Google with respect to MySQL. Had an open development model existed when we started more of our work would have been done in the open, we would have made fewer mistakes and it would have been easier to get our work into mainstream branches. But I think that change is happening. -- Mark Callaghan mdcallag@gmail.com
Hi! On Wed, May 13, 2009 at 10:26:41PM -0700, Vadim Tkachenko wrote:
What do you think about including patches global_trx_ids, binlog event checksums and fixes to crash-proof slaves from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids into tree ?
Speaking in general (and not about these particular patches), my personal opinion is that right now the blocker to including patches is lack of infrastructure to test both the patch and how it applied to maria tree. Me and Kristian have been trying to sort this out last week - added several buildbot slaves and a staging branch to do testing. We're not fully done with this yet [1]. Once we've sorted this out, it will be possible to make changes to the tree more easily. [1]. The problem I'm looking at is that for some reason it takes unreasonable amounts of time and bandwidth to do any operations on the staging trees. I have no idea why, perhaps we're hitting a bug in bazaar's handling of stacked branches. Maybe we should switch to non-stacked ones. Once that is sorted out, there is a mysql-test run failure I'd prefer to see fixed before we add more changes. BR Sergey -- Sergey Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog
Sergey Petrunya <psergey@askmonty.org> writes:
[1]. The problem I'm looking at is that for some reason it takes unreasonable amounts of time and bandwidth to do any operations on the staging trees. I have no idea why, perhaps we're hitting a bug in bazaar's handling of stacked branches. Maybe we should switch to non-stacked ones. Once that is sorted out, there is a mysql-test run failure I'd prefer to see fixed before we add more changes.
Hm. I just tested locally, and these are the times I see: $ time bzr branch --no-tree 'lp:~maria-captains/maria/maria-6.0-testing1' Branched 3040 revision(s). 1.63user 0.10system 0:10.52elapsed 16%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+176outputs (0major+23387minor)pagefaults 0swaps $ cd maria-6.0-testing1/ $ time bzr co 65.67user 1.86system 1:28.47elapsed 76%CPU (0avgtext+0avgdata 0maxresident)k 281400inputs+345344outputs (1major+93710minor)pagefaults 0swaps $ echo extra line >> README $ time bzr ci -m "test commit" Committing to: /home/knielsen/devel/repo/maria-6.0-testing1/ modified README Committed revision 3041. Skipping None: allow_remote_options not set for mysql plugin. 4.00user 0.28system 0:04.46elapsed 95%CPU (0avgtext+0avgdata 0maxresident)k 440inputs+5728outputs (1major+25187minor)pagefaults 0swaps $ time bzr push --stacked-on bzr+ssh://bazaar.launchpad.net/~maria-captains/maria/6.0 bzr+ssh://bazaar.launchpad.net/~knielsen/maria/tmp-test-launchpad-speed4 Created new stacked branch referring to bzr+ssh://bazaar.launchpad.net/~maria-captains/maria/6.0. 1.68user 0.11system 0:19.05elapsed 9%CPU (0avgtext+0avgdata 0maxresident)k 56inputs+32outputs (0major+17078minor)pagefaults 0swaps So 10 seconds to branch, 4 to commit, and 19 to push to launchpad trees. So I wonder what is different at your end to make it slow. Of course, for 6.0 trees it is important to use the --stacked-on option, but since you were the one I learned that from that is unlikely to be your problem. No idea, sorry... - Kristian.
Kristian Nielsen <knielsen@knielsen-hq.org> writes:
No idea, sorry...
Hm, I just remembered something odd on the 6.0 Launchpad tree: https://code.launchpad.net/~maria-captains/maria/6.0 Repository format: Packs containing knits without subtree support This is different from our other repositories, and seems to be an older format. Though I am 99% sure I upgraded the 6.0 tree at the same time as the other trees. So maybe this is what is causing your problems? Though in that case it is still strange that I do not have any problems. Feel free to upgrade the lp:maria/6.0 tree to repository version 1.9 to see if it solves your problem. Also which version of bzr are you using? Here is my: Bazaar (bzr) 1.14.1 Python interpreter: /usr/bin/python 2.5.2 Python standard library: /usr/lib/python2.5 bzrlib: /usr/lib/python2.5/site-packages/bzrlib Bazaar configuration: /home/knielsen/.bazaar Bazaar log file: /home/knielsen/.bzr.log Copyright 2005, 2006, 2007, 2008, 2009 Canonical Ltd. http://bazaar-vcs.org/ bzr comes with ABSOLUTELY NO WARRANTY. bzr is free software, and you may use, modify and redistribute it under the terms of the GNU General Public License version 2 or later. - Kristian.
Hi!
"Vadim" == Vadim Tkachenko <vadim@percona.com> writes:
Vadim> Hello, Vadim> What do you think about including patches Vadim> global_trx_ids, binlog event checksums and fixes to crash-proof slaves Vadim> from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids Vadim> into tree ? During the incoming months we will look at all publicly available patches, included the above, to decide what makes sense to do with them. Basicly there are 3 options: - Review, fix (if necessary) and add the patch to MariaDB 5.1 - Review, fix (if necessary) and add the patch to MariaDB 5.4 Disregard the patch for now, but put a reference to it in worklog with a comment why we disregarded the patch and also, if it makes sense, what needs to be done for us to accept the patch. Regards, Monty
Hi!
"Vadim" == Vadim Tkachenko <vadim@percona.com> writes:
Vadim> Hello, Vadim> What do you think about including patches Vadim> global_trx_ids, binlog event checksums and fixes to crash-proof slaves Vadim> from https://code.launchpad.net/~jtolmer/mysql-server/global-trx-ids Vadim> into tree ? I just created a worklog for this to ensure we don't forget about it. http://askmonty.org/worklog/Server-BackLog/?tid=31 Regards, Monty
participants (6)
-
Kristian Nielsen -
MARK CALLAGHAN -
Michael Widenius -
Paul McCullagh -
Sergey Petrunya -
Vadim Tkachenko