y incorrect (uses the wrong barriers), but
> it happens to generate the correct assembler code on x86 due to its rather
> strong memory ordering.
>
> It is probably still incorrect for some other architectures (eg. multi-core
> ARM).
What about POWER? It's officiall
many problems with
> our debian packaging - all these fixes looks right, and still need
> twenty iterations to actually be right.
What's the issue with oqgraph building? Is it something boost related? I
think I recall some issues with some Boost versions (perhaps only on
pp64le) but I cou
ally pretty easy to spot using perf on
POWER, and my suspicion is that the 10% perf difference is made up of a
couple of things like this and shouldn't be tooo hard to claw back.
--
Stewart Smith
pgpyFlnyChvY9.pgp
Description: PGP signature
___
Ma
t to participate (depending on
installation method).
Another solution is to have feedback plugin as package that you just
install and it gets enabled (by putting config in /etc/mariadb.d/ or
something).
--
Stewart Smith
pgpz1635kfW6T.pgp
Description: PGP signature
rg writes:
> I want to link an external library for a language extension. After several
> hours of trying to tell cmake (^%$@*&) to pick it up for linking the
> library would still not be built with.
did you remove the CMakeCache.txt file?
CMake is garbage like that.
--
haven't looked at the
produced code).
Personally, looks good and I'm tempted to pull it into Percona Server.
--
Stewart Smith
pgpDSwJn4niGi.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to
ounds like a bug we fixed in Percona Server a while ago. filtered
replication is kinda awful and when we started to poke at it, it pretty
much never worked properly. It may be that this is not specific to this
replication patch.
--
Stewart Smith
pgpRgqk2I7BNH.pgp
Description: PGP signature
___
he legacy format,
that would be a very clear break in any on-disk compatibility with MySQL.
--
Stewart Smith
pgpaeGCzt7XsK.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.lau
and easy implement new features, example in json
> we can add a new item in object just put ", 'new item':'some value'",
> i don
Schemaless isn't necessarily a feature - protobuf has the benefit of
enforcing a schema (XML would work too... parsing time was a conc
x27;s do it next month, shouldn't take more than a
> couple of weeks".
>
> Still, never happened.
Happened in Drizzle, IIRC it was about 6 months of work for the bulk of
it.
--
Stewart Smith
pgpWQvUN1Nuft.pgp
Description: PGP signature
__
t.
> humm i read something about protobuf, but is it human readable?
You can outptu it as human readable fairly easily, it's otherwise a
compact binary format that's generally much quicker to parse than human
readable.
k
--
Stewart Smith
pgpo
this though. It appears there's some code
in MariaDB to construct server data structures from a SQL statement
though, so perhaps that's more usable.
--
Stewart Smith
pgpYbrwEa6brw.pgp
Description: PGP signature
___
Mailing list: h
Roberto Spadim writes:
> humm interesting, does mariadb have the same bugs of mysql? or it's updated
> like drizzle?
Same limitations as MySQL due to FRM file format.
--
Stewart Smith
pgpCLC5Ocnmjg.pgp
Description: PGP signature
___
M
ept in memory, and going to that many is going to mean a *big* chunk of
memory that will be always used. It's much better to do the join and be
done with it, it'll end up being a lot more efficient.
--
Stewart Smith
pgpEsXGcVUL9Q.pgp
Description: PGP signature
___
. If it's greater than max_heap_table_size it'll be converted to MyISAM/Maria.
--
Stewart Smith
pgpyXmRXcwKPo.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscrib
les are first materialised into a MEMORY or
MyISAM (or, with MariaDB, possibly Aria) table and then the queries run
on them. There is no INFORMATION_SCHEMA storage engine.
--
Stewart Smith
pgphRmHaVJtpo.pgp
Description: PGP signature
___
Mailing list:
t not being a giant hole into your database server.
--
Stewart Smith
pgp2UARLQ5VUk.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe : https://launchp
ecovery time for the database.
--
Stewart Smith
pgptnNk0rmDe2.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe : https://launchpad.net/~maria-deve
te buffer. I could never reconcile myself with this.
Mind you, it performs really quite well due to the reduced fsync()s and
(at least in Drizzle code) was rather trivial to implement. Big bang for
buck :)
--
Stewart Smith
pgp5gG0zbzgFF.pgp
Description: P
or all
replication up to that point to be committed, and then do the MyISAM
operation. (as there's no reason why other replication couldn't be
applied at the same time... after all, MyISAM isn't crash safe so would
never be consistent with InnoDB txns).
--
Stewart Smith
pgptrrWThqpI3
NDB, both strategies saw performance improvement.
--
Stewart Smith
pgp2LHNSUj8YM.pgp
Description: PGP signature
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe : https://lau
Kristian Nielsen writes:
> Stewart Smith writes:
>
>> Idea: if txn touches more than one database, record as metadata the
>> position of other db, so slave just has to wait for at least that id to
>> be applied in other database before applying this txn in this
>
Kristian Nielsen writes:
> Stewart Smith writes:
>
>> Kristian Nielsen writes:
>>> https://mariadb.atlassian.net/browse/TODO-171
>>
>> this seems to be private, and I can't view even having created a JIRA
>> account.
>
> Oops :-( Very sorr
Sergei Golubchik writes:
> On Aug 10, Stewart Smith wrote:
>>
>> > Suppose we instead make the decision on the _master_ on how to split
>> > up the replication stream for parallel apply. Instead of a single
>> > server-id for the master, we can have multipl
der (though sequences may be
> interleaved differently) - so changing masters is simpler, all servers in the
> hierarchy have the same view.
How would this work with a cross database transaction that depends on
previous transactions in both database?
his isn't necessarily the case,
different GCC versions etc do catch different things. Every time
somebody suggests building on a subset, we find another issue like this
(early) soo... I wonder if you tempt fate the same way we do :)
--
Stewart Smith
pgpyRUclz
ck the SQL syntax but also if all mentioned tables, columns etc.
> exists in the given database.
This (of course) won't always work. ALTER TABLE ADD COLUMN and then
mention that column.
As an alternative, maybe start up a local server and load data in it to
te
tores blobs sep
from main table, gives http interface.
--
Stewart Smith
___
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
this year entitled "Fixed in Drizzle"
and it's going to require a bunch of research as we didn't always keep a
singular comprehensive list (especially non user-visible things).
--
Stewart Smith
___
Mailing list: https://launchpad.net/
Real simple patch, makes the test run much faster if using a real
storage device without libeatmydata.
patch should be appliable (just change the path)
revno: 2477
committer: Stewart Smith
branch nick: update-innobase
timestamp: Tue
way things go.
Better to be able to bisect back to the problem then go "somewhere in
this 600kb diff".
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe
o the client?
> Commit or rollback was done before a response was sent to the client.
One thing this means is that you have to do stupid things managing
memory with BLOBS... but that's just excessive memory usage rather than errors.
--
Stewart Smith
epends on what the correct behaviour of auto_increment
columns is meant to be.
For comparison, it looks as though NDB has taken the "MyISAM is correct"
path and done the update (I read source, didn't actually try it :)
To me, updating the aut
ts for DDL (i.e. anybody that isn't
you, at the same version).
--
Stewart Smith
___
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
serve as examples of how to (or not to) write bits...
--
Stewart Smith
___
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
plexity...
For basic DML, you can check out CSV for probably the most basic yet
functional implementation.
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe
r alive" at most:
a) connects to socket
b) runs "SELECT 1;"
and none of them actually check the storage engine you're using is okay.
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-developers
Post to :
to quilt, but with the added bonus of
each "patch" is revision controlled.
It also has the benefit of each rebase on a new version uses bzr merge
algorithms, so it should mean less work.
Check out "Maintaining a series of patches against software from
tarb
sh to
launchpad and link the branch with the task being worked on (we use
blueprints on launchpad, which has a specific field for this).
you could also use 'bzr send' type things in the commit mails, so you
could even apply the commit mails and merges
at introduces cursors needs
> architecture work and prototyping. I fear we will end up with
> 3 disparate storage engine APIs in Drizzle, MySQL and MariaDB.
The current aim is to remove things that obviously aren't related to a
cursor and move them up.
e.g. create table is
ameters should take care of the only
retreiving/setting certain columns. Although that still leaves optimiser
things... which is another project altogether :)
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-developers
Post to : m
On Wed, Jul 08, 2009 at 03:53:15AM -0400, Peter Benjamin Volk wrote:
> Stewart Smith wrote:
> > The idea being to move the handler to be a cursor on a table, with
> > actions not pertaining that to reside in StorageEngine (e.g. DDL).
> >
> That sounds like an interesting
n the same way?
--
Stewart Smith
___
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
could be useful here too - linking back to MySQL bug tracker or Drizzle
bug tracker (launchpad).
--
Stewart Smith
___
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
atomics for xtradb by default.
We have some headers in drizzle so we get atomic ops through the
SunStudio provided way as well (or gcc). may want to look it up for
better Solaris perf.
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-develo
Antony mentioned on IRC about CMake changes that let you not use
config.js and other nasty things when building on Win32 (and Win64).
These are all just in the MySQL Cluster branch - namely 7.0.
Who knows when this gets back into any mainline branch though...
--
Stewart Smith
he gcc builtin.
--
Stewart Smith
___
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
and merged over the pbxt files that already exist (in the MariaDB and
> Drizzle/PBXT trees)?
That may not be the best... probably lots of conflicts on files and
having to work out which is right (and conflicts on directory paths
too).
best to extract patches you want, commit upst
automatically tests a PBXT commit against current Drizzle and MariaDB.
Hope this helps
--
Stewart Smith
___
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe : https://launchpad.net/~
f seconds\n", do_test (repetitions, speed_loop));
printf ("Testing loop32 ");
fflush (stdout);
printf (" done, %10.3f seconds\n", do_test (repetitions, speed_loop32));
printf ("Testing loop64 ");
fflush (stdout);
printf (" done, %10.3f second
50 matches
Mail list logo
