[MonetDB-users] MonetDB scalability
Hi,
I'm currently evaluating MonetDB for business reporting, and looking for information on scalability, and I can't seem to find any information on clustering options.
Is there any support in MonetDB for clustering across multiple physical servers to improve performance?
Thanks, Chris
Chris wrote:
Hi,
Hi Chris,
I'm currently evaluating MonetDB for business reporting, and looking for information on scalability, and I can't seem to find any information on clustering options.
Is there any support in MonetDB for clustering across multiple physical servers to improve performance?
Nope, that is not supported yet in the current releases. We look forward towards the requirements in terms of performance needed, e.g. schema/queries, which are hard to handle for the current version.
regards, Martin
Thanks, Chris
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
Hi Martin,
Well, nothing concrete yet, just putting together a typical, boring document to fling at management.
Even without clustering, MonetDB has beaten out the other DB options we've tried by a huge margin.
If I come across any specific performance concerns, I'll be sure to let you know.
Thanks, Chris
On 07/07/2010, at 4:19 PM, Martin Kersten wrote:
Chris wrote:
Hi,
Hi Chris,
I'm currently evaluating MonetDB for business reporting, and looking for information on scalability, and I can't seem to find any information on clustering options.
Is there any support in MonetDB for clustering across multiple physical servers to improve performance?
Nope, that is not supported yet in the current releases. We look forward towards the requirements in terms of performance needed, e.g. schema/queries, which are hard to handle for the current version.
regards, Martin
Thanks, Chris
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
If you are looking at alternatives check out the Infobright database. I'm in the process of switching my infrastructure over to that database, it uses considerably less memory and doesn't crash all the time.
On Wed, Jul 7, 2010 at 8:17 PM, Chris gaffer@gaffertaped.com wrote:
Hi Martin,
Well, nothing concrete yet, just putting together a typical, boring document to fling at management.
Even without clustering, MonetDB has beaten out the other DB options we've tried by a huge margin.
If I come across any specific performance concerns, I'll be sure to let you know.
Thanks, Chris
On 07/07/2010, at 4:19 PM, Martin Kersten wrote:
Chris wrote:
Hi,
Hi Chris,
I'm currently evaluating MonetDB for business reporting, and looking for
information on scalability, and I can't seem to find any information on clustering options.
Is there any support in MonetDB for clustering across multiple physical
servers to improve performance?
Nope, that is not supported yet in the current releases. We look forward towards the requirements in terms of performance needed, e.g. schema/queries, which are hard to handle for the current version.
regards, Martin
Thanks, Chris
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
On Thu, Jul 8, 2010 at 2:09 PM, Matthew Jones mat@matburt.net wrote:
If you are looking at alternatives check out the Infobright database. I'm in the process of switching my infrastructure over to that database, it uses considerably less memory and doesn't crash all the time.
Don't you find the lack of ALTER TABLE a bit limiting for a production database?
(I assume you are talking about the GPL v2 version.)
Though I too am frustrated with the instability.
m
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
Op 08-07-10 20:18, Mark Bucciarelli schreef:
Though I too am frustrated with the instability.
Typically what kind of things do you see regarding to instability? I'm currently running multiple webservices with a bunch of data and an open readonly SQL interface.
The only crashes I observe are related: 1) silly SQL (thats basically why the webinterface is public, to catch those bugs) 2) usually some oddness with corrupt input (for example with copy into)
Is your current production workload completely different? Because I have had uptimes exceeding > 4 months, and just to upgrade the software.
Stefan
On Thu, Jul 8, 2010 at 2:53 PM, Stefan de Konink stefan@konink.de wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Typically what kind of things do you see regarding to instability?
Yesterday, it was wrong contraint id in fk violation error (bug #2598, open).
Day before that, it was a select that reliably crashed db (#2583, which Niels fixed prompty, thank you).
Dropping a contraint corrupts DB (#2342). Reported back in March, with script that produces the behavior reliably, bug is still open.
Dropping a column corrupts DB (#2341). Same as above.
m
On Thu, Jul 8, 2010 at 3:05 PM, Mark Bucciarelli mkbucc@gmail.com wrote:
On Thu, Jul 8, 2010 at 2:53 PM, Stefan de Konink stefan@konink.de wrote: -----BEGIN PGP SIGNED MESSAGE-----
Typically what kind of things do you see regarding to instability?
Other things I've noticed but haven't reported:
mclient shell redirection looks strange to me. try this:
$ mclient -l sql -d test -u test 1>1.out 2>2.out
i've seen other unix utils that request a password navigate this successfully.
i've seen behavior change depending on the case of the text in the SQL. i don't remember exactly what now, as i just worked around it.
i've also had trouble with the server crashing when i tried setting ulimit's. sjoerd wrote great stuff, and i carefully followed it, but it didn't work for me.
finally, from reading the code, i've seen that strdup() return values are often (usually?) ignored. there was even one case where fopen() return value was ignored, someone fixed that after I posted to the dev list.
shared library handling makes some assumptions that aren't valid on all platforms.
mx was segfaulting before sjoerd applied a variant of my patch (bug #2330).
That's all I can think of at the moment.
m
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
Op 08-07-10 21:19, Mark Bucciarelli schreef:
finally, from reading the code, i've seen that strdup() return values are often (usually?) ignored. there was even one case where fopen() return value was ignored, someone fixed that after I posted to the dev list.
Same for sprintf, while a strlen usually is ran after it... (I know because I wrote a patch to fix them all, and pointed out the missing error checking.)
Stefan
On Thu, Jul 08, 2010 at 03:05:35PM -0400, Mark Bucciarelli wrote:
On Thu, Jul 8, 2010 at 2:53 PM, Stefan de Konink stefan@konink.de wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Typically what kind of things do you see regarding to instability?
Yesterday, it was wrong contraint id in fk violation error (bug #2598, open).
Just fixed
Day before that, it was a select that reliably crashed db (#2583, which Niels fixed prompty, thank you).
Dropping a contraint corrupts DB (#2342). Reported back in March, with script that produces the behavior reliably, bug is still open.
Dropping a column corrupts DB (#2341). Same as above.
Both were fixed some time ago and tests were added by sjoerd. These run fine, ie I closed them.
Niels
m
This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ MonetDB-users mailing list MonetDB-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-users
On Thu, Jul 8, 2010 at 6:28 PM, Niels Nes Niels.Nes@cwi.nl wrote:
On Thu, Jul 08, 2010 at 03:05:35PM -0400, Mark Bucciarelli wrote:
Yesterday, it was wrong contraint id in fk violation error (bug #2598, open).
Just fixed
Thanks.
Dropping a column corrupts DB (#2341). Same as above.
Both were fixed some time ago and tests were added by sjoerd. These run fine, ie I closed them.
I had no idea.
Today's entry:
#2602 WHERE clause with AND NOT is broken
m
participants (6)
-
Chris -
Mark Bucciarelli -
Martin Kersten -
Matthew Jones -
Niels Nes -
Stefan de Konink