esults in a faster way.
PostgreSQL 8.3 contains great improvements in this area, you can simply
start the selects from concurrent connections, and the backend will
synchronize the scans.
Regards,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Softwar
mple configuration, there are others on different machines
where it turns out that forcing index usage leads to faster queries, and
less overall ressource consumption. (Or, at least, faster delivery of
the first part of the result so the application can begin to process it
asynchroneously).
Thank
Hi, Steinar,
Steinar H. Gunderson wrote:
> On Wed, Nov 22, 2006 at 11:17:23AM +0100, Markus Schaber wrote:
>> The Backend allocates gigs of memory (we've set sort_mem to 1 gig), and
>> then starts spilling out more Gigs of temporary data to the disk.
>
> How much RAM
irrelevant.
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
---(end of broadcast)---
TIP 9: In vers
s your free space map setting, the statistics targets, and other
config options tuned to fit your environment?
- Maybe a VACUUM FULL or a CLUSTER command may help you.
> for a single select with one field in one resultset, is 0.86 seconds normal?
That depends on the circumstances.
Markus
--
M
on, and if not, do a "time dd
if=/your/device of=/dev/null bs=1M" on the partition.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
--
Hi, Craig,
Craig A. James wrote:
>> Would dblink() help in any way?
>
> It might if perl wasn't so damned good at this. ;-)
You know that you can use Perl inside PostgreS via plperl?
HTH,
Markus
---(end of broadcast)---
TIP 3: Have you checked o
index updates cause both CPU load, and random disk access (which is
slow by nature).
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
-
t versions of the same tuple), and so an index
change is not necessary when both versions fit on the same page.
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ff
a new row version in the same heap page,
> the index must be updated anyway.
AFAICS only, when the index covers (directly or via function) a column
that's actually changed.
Changing columns the index does not depend on should not need any write
access to that index.
Correct me if I'm
s). So it may be necessary to
re-check the condition with the real data, using the lossy index for
preselection.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www
out making your perl program writing the COPY
statement as text, and piping it into psql.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
--
n equal disks? And in the same areas of those disks? Some
current disks tend to drop down their speed at the "end" of the LBA
address space drastically.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development
7; numbers to be queried per group (all 40 groups
> are displayed on the same page) and so constant aggregates over the
> entire table would be a nightmare.
That sounds just like a case for GROUP BY and a materialized view.
Markus
--
Markus Schaber | Logical Tracking&Tracing Inte
act, the SERIAL type does nothing but defining a sequence, and then
use nextval('sequencename') as default.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Eur
ing real row counts etc.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
---(end of broadcast)
ll think that using a PL in the backend might be more performant
than having an external client, alone being the SPI interface more
efficient compared to the network serialization for external applications.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. |
p by", "order by", "distinct on" and hand-written functions
and aggregates (like first() or best()) may help.
You could combine all relevant columns into an user-defined compund
type, then group by entity, and have a self-defined aggregate generate
the accumulated tu
Hi, Uwcssa,
uwcssa wrote:
>
> Please unsubscribe me! Thank you!
Sorry, but we (the list members) are unable do that, we have no
adminstrative power on the list. :-(
> Also, it would be better to have a message foot saying how to unsubscribe.
List unsubscribe information is contained in t
if the query plan stays equal.
Do you look up the definition of cost for Informix? Have you made shure
that they're comparable?
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Eur
ovacuum to kick in are configurable, see the
docs.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
--
the
desired result?
We're using this way for some 3rd-party databases we have to process
in-house.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwar
Hi, Marc,
Marc Morin wrote:
>> I wonder whether there is a way to use table partitioning to
>> make the insert pattern more localized? We'd need to know a
>> lot more about your insertion patterns to guess how, though.
>
> We're doing partitioning as well.
And is constraint exclusion set
e new multi-row INSERT
to further improve performance, but I doubt that you're CPU bound.
The only way to "really" get over the sync limit is to have (at least)
the WAL on a battery backed ram / SSD media that has no "spinning disk"
physical limit, or abandon crash safety by t
a running top to
multi-cpu mode, pressing "1" again switches back to accumulation.
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
-
little, but I don't know whether it's worth the effort, given that
bitmap indidex scans exist.
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.
as the
application which has more informations about semantics and algorithms.
Everything else would need a crystal ball device :-)
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents
do you have a concrete
problem? In latter case, you could post your details here, and we'll see
whether we can help.
Btw, I'm not related with Bizgres in any way. :-)
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Softwar
I hope that this lengthy mail is enlightening, if not, don't hesitate to
ask.
Thanks for your patience,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatent
Hi, Luke,
Luke Lonergan wrote:
>> I thought that posix_fadvise() with POSIX_FADV_WILLNEED was exactly
>> meant for this purpose?
>
> This is a good idea - I wasn't aware that this was possible.
This possibility was the reason for me to propose it. :-)
> We'll do some testing and see if it work
Hi, Luke,
Luke Lonergan wrote:
>> Do you think that adding some posix_fadvise() calls to the backend to
>> pre-fetch some blocks into the OS cache asynchroneously could improve
>> that situation?
>
> Nope - this requires true multi-threading of the I/O, there need to be
> multiple seek operation
ily be #define'd out on
platforms that don't support it.
Combine this with the Linux Kernel I/O Scheduler patches (readahead
improvements) that were discussed here in summer...
Regards,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Developmen
I don't have much more ideas what the problem could be.
Can you try to do some profiling (e. G. with statement logging) to see
what specific statements are the one that cause high cpu load?
Are there other differences (besides the PostgreSQL version) between the
two installations? (Kernel,
ot broken or misconfigured, always reconnect?
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
---(end of broadcast)-
is one of the more debated points in the PostgreSQL way of MVCC
implementation.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarep
at them as comments, and uses the defaults instead. You should make
shure that you use "real" settings in your config.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Eur
h kill
performance.
> with 2 Go RAM and good SCSI disks.
For 2 Gigs of ram, you should cut down the number of concurrent backends.
Does your machine go into swap?
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight
command in their
benchmarks.
Realistic in-production workloads don't have so much create table
commands, I think.
Wondering,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
this behaviour? it'is lovely...:)
>
> Tom , can you explain why?...
It is because the transaction IDs are global per cluster.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! ww
ty might help you identify the
> culprit(s).
Another possibility might be an outstanding two-phase-commit transaction.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU!
with a bloat value <1 and a
negative "wasted space" - is this due to the pseudo nature of them?
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! ww
for the whole
> database)
For 7.4, you'll need a script to do that (current versions have improved
in this area). You might recycle the idea from the pgsql-sql list some
days ago:
http://archives.postgresql.org/pgsql-sql/2006-08/msg00184.php
Simply use the meta tables to get a list of al
nk that vacuum during working hours
puts too much load on your server, there are options to tweak that, at
least in 8.1.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.n
Unfortunately, we didn't keep the logs of VACUUM ANALYZE, so I
> can't be sure :/
AFAIK, the warning is also output on the psql command line.
HTH,
Markus
[1] We once had an index that was about 100 times larger before REINDEX.
--
Markus Schaber | Logical Tracking&Tracing Internatio
em.
It also might make sense to issue a CLUSTER instead (which combines the
effects of VACUUM FULL, REINDEX and physically reordering the data).
When the free_space_map is to low, VACUUM ANALYZE should have told you
via a warning (at least, if your logging is set appropriately).
HTH,
Markus
--
Markus
Hi, Jeff & all,
Jeff Davis wrote:
> (2) You have a long-running transaction that never completed for some
> strange reason.
I just asked myself whether a 2-phase-commit transaction that was
prepared, but never committed, can block vacuuming and TID recycling.
Markus
--
Mark
debian with their multi-cluster
multi-version script wrapper magic.
Don't mind.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
-
elf, that I had installed the
correct postgresql and psql versions, but accidentally connected to a
different database installation due to strange environment and script
settings...
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Softwa
ll, actually, the controller should not reorder over write barriers.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---
ewall or antivirus network filtering software.
HTH,
Marku
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)---
TIP 6: explain analyze is your friend
rying to estimate future usage patterns).
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)-
that. Thank you.
It seems that you, by accident, hit the wrong list with your question.
But, as you're here, why don't you migrate to PostgreSQL instead?
Have a nice day,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development G
settings might help you out.
As far as I understand, this will only help for concurrent inserts by
different clients, dealing throughput for latency. Please correct me if
I'm wrong.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Deve
ound writer, stats collector or autovacuum, depending on your
version and configuration.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
g the kernel.
As stated on the
http://www.postgresql.org/docs/8.1/interactive/kernel-resources.html
page, those values can be changed via sysctl or echoing values into
/proc, under linux at least.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Sof
Hi, Scott and Hale,
Scott Marlowe wrote:
> Make sure analyze has been run and that the statistics are fairly
> accurate.
It might also help to increase the statistics_target on the column in
question.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
se results for low concurrency situations).
Different fsync method settings can also make a difference (I presume
that syncing was enabled).
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against so
'm used to with the Adaptec SCSI junk.
Well, for sequential reading, you should be able to get double drive
speed on a 2-disk mirror with a good controller, as it can balance the
reads among the drives.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf.
[1] We once had such a problem because an ill-compiled kernel having USB
verbose logging on...
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
--
Hi, Larry,
Hi, Sven,
Sven Geisler wrote:
> You can increase wal_buffers, checkpoint_segments and checkpoint_timeout
> much higher.
You also should increase the free space map settings, it must be large
enough to cope with your weekly bunch.
Markus
--
Markus Schaber | Logical Tr
gh and/or don't have your FSM
> setting high enough.
Depending on the PostgreSQL version, it might also be that he suffers
from index bloat. He might look into the manual pages about REINDEX for
a description.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing Inter
'll definitely fiddle with that in the coming
> tests as well.
How many parallel transactions do you have?
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU!
e system.)
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
is is for plpgsql, pljava and plpython?
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)
the affected-row count
> on the last rule which gives Hibernate problems.
This could be considered a PostgreSQL bug - maybe you should discuss
this on the appropriate list (general, hackers)?
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. |
ble, to have
current statistics.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)-
dropping and recreation.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)---
PostgreSQL) we use the one schema per customer
paradigm quite successfully.
> My $0.02 (not worth what it was)
Oh, I think the're at least $0.03 cents worth. :-)
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight again
d from listing all schemas (or even
all databases in the same user), regardless whether he/she has access
rights.
But it is not always acceptable that a customer knows which other
customers one has.
This forces the use of the "one cluster per customer" paradigm.
Thanks,
Markus
--
Markus S
dx ON purchase_order_items
(expected_quantity) WHERE expected_quantity > 0;
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---
s is pretty clearly in the 1%
> it can't handle.
Maybe your free space map is configured to small, can you watch out for
log messages telling to increase it?
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight ag
when leaving it open while having lunch...
Some older JDBC driver versions had the bug that they always had an open
transaction, thus an application server having some pooled connections
lingering around could block vacuum forever.
Markus
--
Markus Schaber | Logical Tracking&Tracing Inte
Hi, Tim,
Seems I sent my message to fast, cut in middle of a sencence:
Markus Schaber wrote:
>> A pg_dump/pg_restore cycle reduced the total
>> database size from 81G to 36G.
> If you still have the original database around,
... can you check wether VACUUM FULL and REINDEX a
e, and increase the WAL size /
checkpoint segments.
When most of the restore time was spent in index creation, increase the
sort mem / maintainance work mem settings.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fig
le.
Most of the cases when we had database bloat despite running autovacuum,
it was due to a low free_space_map setting.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosof
2 G.
This sounds like you want to have one postgresql backend per apache
frontend.
Did you try running pgpool on the Apache machine, and have only a few
(hundred) connections to the backend?
Maybe http://en.wikipedia.org/wiki/Memcached could be helpful, too.
Markus
--
Markus Schaber | Lo
atabase is only 58M it's a read only DB and will lasts only for a
> month.
I guess it is a simple table with a single PK (some subscription numer)
- no joins or other things.
For this cases, a special non-RDBMS like MySQL, SQLite, or even some
hancrafted thingy may give you better resu
a, most Datasource implementations (e. G. the JBoss one)
allow to specify SQL statements that are executed on connection init.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www
and input
file to different spindles won't help you much.
HTH
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---
ctions,
and then give them at least 30 megs of RAM each.
This should also cut down the connection creation overhead.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! ww
Hi, Mario,
Mario Splivalo wrote:
> This helps also. I don't get sequential scans any more. I'd like a tip
> on how to set 'enable_nestloop = off' trough JDBC?
statement.execute("SET enable_nestloop TO off"); should do.
HTH,
Markus
--
Markus Schaber | Logic
aps a better solution would be to cache the
> result of the estimator function.
Sophisticated estimator functions are free to use the pg_statistics
views for their row count estimation.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Dev
E. G. for generate_series we return ($2-$1)/$3, and for some functions
even constant estimates will be good enough.
> - please execute me and store my results in a temporary storage,
> count the rows returned, and plan the outer query accordingly
That's an interesting idea.
Hi, Bruce,
Markus Schaber wrote:
>>>It does not find as much liers as the script above, but it is less
>>Why does it find fewer liers?
>
> It won't find liers that have a small "lie-queue-length" so their
> internal buffers get full so they have to bloc
), but not
syncs/write rate.
Both of them can be captured by the other script, but not by my tool.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatent
SB/FireWire to IDE case, and an older linux cryptoloop implementations,
IIRC).
If you're interested, I can dig for the C source...
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software
e estimator
functions have an implicit return parameter of int8. Parameters may be
NULL when they are not known at query planning time.
What do you think about this idea?
The same scheme could be used to add a CPUCOST_ESTIMATOR to expensive
functions.
HTH,
Markus
--
Markus Schaber | Logical Track
ur indices using REINDEX command.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)-
Additonally, the "free_space_map" setting has to be high enough, it has
to cover enough space to put in all pages that get dead rows between two
vacuum runs.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight
ter".flatten_array = streets.link_id)
(4 rows)
Currently, we're planning to use the array flattening approach, but are
there any plans to enhance the query planner for the direct ARRAY approach?
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. I
g with commit_delay and
commit_siblins may improve your situation.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of
will try to access a
not-completely-restored wal file.
HTH,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---
hink that this is an easy, understandable and useful interpretation
of posix_fadvise() hints.
Are there any rough estimates when this will get into mainline kernel
(if you intend to submit)?
Thanks,
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. |
to a read-only table for now.
Are they capable to index custom datatypes like the PostGIS geometries
that use the GIST mechanism? This could probably speed up our Geo
Databases for Map rendering, containing static data that is updated
approx. 2 times per year.
Markus
--
Markus Schaber | Logical Tr
oject-du-jour are
> "dumbed down" to the mysql featureset. (And not just mysql, but
> mysql-lowest-common-factors, which means myisam etc)
Well, most of those projects don't need a database, they need a bunch of
tables and a lock.
Heck, they even use client-side SELECT-loops
g extra delays.
Well, if you read the documentation, you will see that it will only wait
if there are at least commit_siblings other transactions active. So when
Bacula serializes access, there will be no delays, as there is only a
single transaction alive.
HTH
Markus
--
Markus Schaber | Lo
s a crude way of partially compensating for this basic problem.
I totall agree with this, it's just what we did here from time to time. :-)
Hmm, how does effective_cach_size correspond with it? Shouldn't a high
effective_cache_size have a similar effect?
Thanks,
Markus
--
Markus Schaber |
at has successfully finished a commit
can be shure their data is on the platters.[1]
HTH,
Markus
[1] As long as the platters don't lie, but that's another subject.
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against so
it infrastructure for them.
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii.org www.nosoftwarepatents.org
---(end of broadcast)---
T
indices on different disks / raid arrays, to parallelize writes.
Btw, I guess you have multiple, concurrent users?
Markus
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in EU! www.ffii
1 - 100 of 183 matches
Mail list logo