PostgreSQL on Linux PC vs MacOS X

[Jim Strickland]

We currently are running a data intensive web service on a Mac using 4D.
The developers of our site are looking at converting this web service to
PostgreSQL. We will have a backup of our three production servers at our
location. The developers are recommending that I purchase a 2GHz Dual
Processor G5 with between 2GB and 4 GB RAM. They say that this
configuration would be able to easily run a copy of all three production
servers. My question is: has anybody had any experience comparing the
performance of PostgreSQL on a G5 Mac versus a PC running Linux? Can
anyone tell me if there are any benefits of running PostgreSQL on one
platform over the other. Anything that can help me make the best
decision would be appreciated.

--
James Strickland - MCP
IT Manager
American Roamer
901-377-8585
http://www.americanroamer.com
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

[Kevin Murphy]

On Nov 3, 2004, at 1:33 PM, Jeff Bohmer wrote:

We use PostgreSQL 7.x on both OS X and Linux. We used to run OS X in
production, but due to numerous problems we switched to Linux. OS X
was not stable at all, especially under load. It was also a poor
performer under load or not.

Did you (or anyone) ever compare performance of PostgreSQL under PPC
Linux running on the G4 or G5?

-Kevin Murphy
---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

[Kevin Barnard]

OS 10.3 IMHO is more stable then 10.2. I haven't us OS X in a
production environment only for development. I have yet to have any
problems with it crashing.

I haven't really run any tests to load it down but that's only because
I never expect to use in production. We have far too many IBM Servers
with battery backed up RAID controllers that I do not see a sudden
switch to any other platform.

If I was a gambling man I would put my money on Linux doing a better
job with postgres, but that's mainly because of the better hardware
options in regard to disks. If your DB is processor heavy the G5 will
most likely out perform x86 processors. If you go with the XRaid I
think all bets are off with regards to dollar for dollar PC/Mac
comparison..
On Wed, 03 Nov 2004 15:04:03 -0600, Jim Strickland
<ji**@americanr oamer.com> wrote:

Well, the whole reason I have asked this question is because my
developer swears by OS X and PostgreSQL. However, I wanted opinions from
other people who have possibly used a similar setup so I can make an
informed decision. I will certainly keep your advice in mind. I guess
the only reason I was asking about the version of OS X and the G5
processor, is because that is all my developer uses and he seems to
think they make a great combination, but that seems to be at odds with
your experience.

Perhaps some others will weigh in with their experiences and I will be
able to make a sound decision. Fortunately there is no great rush to
decide. Thanks for your help.

Jeff Bohmer wrote:

I noticed you ran PostgreSQL on a G4. What version of OS X were you
running? Is it possible the issues you were facing were fixed with
the newer G5 processor?

We were using OS X 10.2 in production. We currently use 10.3 for our
development machines.

I would be shocked if a processor could fix stability issues in an
operating system. As for performance, I cannot say how much better
PostgreSQL runs on a G5 as we don't have any G5s. In terms of
hardware specs, a G4/1.25Ghz should blow away a P3/800. But it didn't
for us, and I think that is because Linux/x86 is much more efficient
than OS X/ppc. I do not expect that to change with a newer ppc
processor.

Since your your developers believe a dual G5 to be plenty, you will
probably get more than enough performance from an XServe G5 and any
comparable 2-way Intel or AMD x86 system. PostgreSQL should handily
outperform 4D. If those systems are in your price range, and
stability isn't a big concern, you should probably go with the OS you
are more familiar with.

- Jeff

Jeff Bohmer wrote:
We use PostgreSQL 7.x on both OS X and Linux. We used to run OS X
in production, but due to numerous problems we switched to Linux. OS
X was not stable at all, especially under load. It was also a poor
performer under load or not.

In my tests, a P3/800, 512MB RAM (100MHz bus) was consistently
faster at all queries than a G4/1.25GHz, 1.5GB RAM (266MHz bus) for
our application. Both machines had single IDE drives.

Another thing to consider is that you can only get ATA drives with
Apple hardware. SCSI is not available from Apple, and SCSI devices
have very poor support under OS X. If a server with ATA drives goes
down at the wrong time, you can lose data. This happened to us with
our production OS X server last year. An extended power outage ran
out the UPS battery, the shutdown script did not stop the server in
time, and we had to restore from an earlier backup. For details on
why this can happen with ATA drives, see this thread:

<http://archives.postgr esql.org/pgsql-general/2003-10/msg01343.php>

Overall, PostgreSQL has been rock solid, very fast, and
headache-free on Linux. A complete change from OS X. Our main
production PostgreSQL server has been up for 234 days now. In that
period, the only downtime for PostgreSQL has been for planned upgrades.

As a side note, we've also had major problems running multi-threaded
servers on OS X which run great (stable and much, much faster) on
Linux.

- Jeff
> We currently are running a data intensive web service on a Mac
> using 4D. The developers of our site are looking at converting this
> web service to PostgreSQL. We will have a backup of our three
> production servers at our location. The developers are recommending
> that I purchase a 2GHz Dual Processor G5 with between 2GB and 4 GB
> RAM. They say that this configuration would be able to easily run a
> copy of all three production servers. My question is: has anybody
> had any experience comparing the performance of PostgreSQL on a G5
> Mac versus a PC running Linux? Can anyone tell me if there are any
> benefits of running PostgreSQL on one platform over the other.
> Anything that can help me make the best decision would be appreciated.
>
> --
> James Strickland - MCP
> IT Manager
> American Roamer
> 901-377-8585
> http://www.americanroamer.com
>
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org

--
James Strickland - MCP
IT Manager
American Roamer
901-377-8585
http://www.americanroamer.com

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ma*******@postg resql.org

[Kevin Barnard]

OS 10.3 IMHO is more stable then 10.2. I haven't us OS X in a
production environment only for development. I have yet to have any
problems with it crashing.

I haven't really run any tests to load it down but that's only because
I never expect to use in production. We have far too many IBM Servers
with battery backed up RAID controllers that I do not see a sudden
switch to any other platform.

If I was a gambling man I would put my money on Linux doing a better
job with postgres, but that's mainly because of the better hardware
options in regard to disks. If your DB is processor heavy the G5 will
most likely out perform x86 processors. If you go with the XRaid I
think all bets are off with regards to dollar for dollar PC/Mac
comparison..
On Wed, 03 Nov 2004 15:04:03 -0600, Jim Strickland
<ji**@americanr oamer.com> wrote:

Well, the whole reason I have asked this question is because my
developer swears by OS X and PostgreSQL. However, I wanted opinions from
other people who have possibly used a similar setup so I can make an
informed decision. I will certainly keep your advice in mind. I guess
the only reason I was asking about the version of OS X and the G5
processor, is because that is all my developer uses and he seems to
think they make a great combination, but that seems to be at odds with
your experience.

Perhaps some others will weigh in with their experiences and I will be
able to make a sound decision. Fortunately there is no great rush to
decide. Thanks for your help.

Jeff Bohmer wrote:

I noticed you ran PostgreSQL on a G4. What version of OS X were you
running? Is it possible the issues you were facing were fixed with
the newer G5 processor?

We were using OS X 10.2 in production. We currently use 10.3 for our
development machines.

I would be shocked if a processor could fix stability issues in an
operating system. As for performance, I cannot say how much better
PostgreSQL runs on a G5 as we don't have any G5s. In terms of
hardware specs, a G4/1.25Ghz should blow away a P3/800. But it didn't
for us, and I think that is because Linux/x86 is much more efficient
than OS X/ppc. I do not expect that to change with a newer ppc
processor.

Since your your developers believe a dual G5 to be plenty, you will
probably get more than enough performance from an XServe G5 and any
comparable 2-way Intel or AMD x86 system. PostgreSQL should handily
outperform 4D. If those systems are in your price range, and
stability isn't a big concern, you should probably go with the OS you
are more familiar with.

- Jeff

Jeff Bohmer wrote:
We use PostgreSQL 7.x on both OS X and Linux. We used to run OS X
in production, but due to numerous problems we switched to Linux. OS
X was not stable at all, especially under load. It was also a poor
performer under load or not.

In my tests, a P3/800, 512MB RAM (100MHz bus) was consistently
faster at all queries than a G4/1.25GHz, 1.5GB RAM (266MHz bus) for
our application. Both machines had single IDE drives.

Another thing to consider is that you can only get ATA drives with
Apple hardware. SCSI is not available from Apple, and SCSI devices
have very poor support under OS X. If a server with ATA drives goes
down at the wrong time, you can lose data. This happened to us with
our production OS X server last year. An extended power outage ran
out the UPS battery, the shutdown script did not stop the server in
time, and we had to restore from an earlier backup. For details on
why this can happen with ATA drives, see this thread:

<http://archives.postgr esql.org/pgsql-general/2003-10/msg01343.php>

Overall, PostgreSQL has been rock solid, very fast, and
headache-free on Linux. A complete change from OS X. Our main
production PostgreSQL server has been up for 234 days now. In that
period, the only downtime for PostgreSQL has been for planned upgrades.

As a side note, we've also had major problems running multi-threaded
servers on OS X which run great (stable and much, much faster) on
Linux.

- Jeff
> We currently are running a data intensive web service on a Mac
> using 4D. The developers of our site are looking at converting this
> web service to PostgreSQL. We will have a backup of our three
> production servers at our location. The developers are recommending
> that I purchase a 2GHz Dual Processor G5 with between 2GB and 4 GB
> RAM. They say that this configuration would be able to easily run a
> copy of all three production servers. My question is: has anybody
> had any experience comparing the performance of PostgreSQL on a G5
> Mac versus a PC running Linux? Can anyone tell me if there are any
> benefits of running PostgreSQL on one platform over the other.
> Anything that can help me make the best decision would be appreciated.
>
> --
> James Strickland - MCP
> IT Manager
> American Roamer
> 901-377-8585
> http://www.americanroamer.com
>
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org

--
James Strickland - MCP
IT Manager
American Roamer
901-377-8585
http://www.americanroamer.com

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to ma*******@postg resql.org

[Jim Crate]

on 11/3/04, Jeff Bohmer <bo****@visionl ink.org> wrote:

We use PostgreSQL 7.x on both OS X and Linux. We used to run OS X in
production, but due to numerous problems we switched to Linux. OS X
was not stable at all, especially under load. It was also a poor
performer under load or not.

In my tests, a P3/800, 512MB RAM (100MHz bus) was consistently faster
at all queries than a G4/1.25GHz, 1.5GB RAM (266MHz bus) for our
application. Both machines had single IDE drives.
In my experience, a G4/1.25GHz computer with standard apple drive was much
faster than the PC (Pentium 2+GHz, don't remember details) we tested running
Linux. Both machines had plenty of RAM, same PostgreSQL settings, etc. The PC
was much slower than the mac running backup/restore (more than 2x slower). The
queries we tested were slower as well. Both machines had IDE drives. I'd think
the Linux box could probably be made to be faster, but it had a long way togo
to even match the G4.

We have had excellent stability on both G4 and G5, MacOS 10.2.x and 10.3.x,
PostgreSQL 7.3.x and 7.4.x. The only time we experienced instability was just
after the G5 was released, the combination of G5, MacOS 10.2.7 and PostgreSQL
7.3.x just didn't work very well. Upgrading the G5 to MacOS 10.3.x and
PostgreSQL 7.4.x brought back the stability we expected and we haven't really
had any problems since.

As a side note, we've also had major problems running multi-threaded
servers on OS X which run great (stable and much, much faster) on
Linux.

Any specific examples? Apache 2.x with PHP 4.3.x has been running well on
various single and dual-cpu MacOS X boxes here.

--
Jim Crate
Deep Sky Technologies, Inc.

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddres sHere" to ma*******@postg resql.org)

[Jeff Bohmer]

>In my experience, a G4/1.25GHz computer with standard apple drive was much

faster than the PC (Pentium 2+GHz, don't remember details) we tested running
Linux. Both machines had plenty of RAM, same PostgreSQL settings,
etc. The PC
was much slower than the mac running backup/restore (more than 2x
slower). The
queries we tested were slower as well. Both machines had IDE
drives. I'd think
the Linux box could probably be made to be faster, but it had a long way to go
to even match the G4.
One possible explanation for your results would be that the Mac IDE
drive lies about write completion while the PC IDE drive does not.
You mention a backup/restore test, which is very write-intensive.
Any system with an IDE drive that lies about write completion is
going to blow away (write performance-wise) a system with an IDE
drive that does not lie about it. Our tests last year were all with
SELECT queries to prevent this factor from skewing our results. (Our
app is read-heavy and we knew we would be getting a good hardware
RAID setup that could handle the writes.)

I do not have the same Apple hardware from a year ago to reproduce my
tests. If I get time in the next week, I can try something on the
same PC (RedHat 9, P3/800) vs. a G4/933, OS X Server 10.2.

We have had excellent stability on both G4 and G5, MacOS 10.2.x and 10.3.x,
PostgreSQL 7.3.x and 7.4.x. The only time we experienced instability was just
after the G5 was released, the combination of G5, MacOS 10.2.7 and PostgreSQL
7.3.x just didn't work very well. Upgrading the G5 to MacOS 10.3.x and
PostgreSQL 7.4.x brought back the stability we expected and we haven't really
had any problems since.

Our primary OS X 10.2 server crashed very frequently. Sometimes more
than once per day. We changed machines and the crashes continued.
Apple HW test on both boxes showed no problems. The vast majority of
these crashes were under moderate load (~120 queries/min). A few
times, reindexing would cause a crash without any other DB activity.
With almost all of these crashes, there were no CrashReporter log
entries. At that point, we felt like we had no recourse but to try
something different (Linux/x86) and haven't looked back.

- Jeff

--

Jeff Bohmer
VisionLink, Inc.
_______________ _______________ ___
303.402.0170 x121
http://www.visionlink.org/
_______________ _______________ ___
People. Tools. Change. Community.

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

[Philippe Schmid]

In my experience, a G4/1.25GHz computer with standard apple drive was
much
faster than the PC (Pentium 2+GHz, don't remember details) we tested
running
Linux. Both machines had plenty of RAM, same PostgreSQL settings,
etc. The PC
was much slower than the mac running backup/restore (more than 2x
slower). The
queries we tested were slower as well. Both machines had IDE drives.
I'd think
the Linux box could probably be made to be faster, but it had a long
way to go
to even match the G4.
One possible explanation for your results would be that the Mac IDE
drive lies about write completion while the PC IDE drive does not. You
mention a backup/restore test, which is very write-intensive. Any
system with an IDE drive that lies about write completion is going to
blow away (write performance-wise) a system with an IDE drive that
does not lie about it. Our tests last year were all with SELECT
queries to prevent this factor from skewing our results. (Our app is
read-heavy and we knew we would be getting a good hardware RAID setup
that could handle the writes.)

I do not have the same Apple hardware from a year ago to reproduce my
tests. If I get time in the next week, I can try something on the
same PC (RedHat 9, P3/800) vs. a G4/933, OS X Server 10.2.

I would say this test is not very meaningful as OS X Server 10.2 is
old, and...not as stable as one could hope, especially if used for
other services.
OS X Server 10.3 is better / much more stable (the "good enough"
depends on what you are doing with it).

Still, I would recommend using the standard OS X 10.3 on a separate box
and tune the kernel params for optimal Postgres use, this is if you
already have one spare Mac for that. Upgrading to the G5 will only be
really usefull if you have very CPU intensive queries. If you are
essentially i/o bound, than faster disks make more sense, and as
previously said, fast disks on a Mac require a third-party SCSI card
and an external RAID or an xRaid. Can be expensive or "cheap"
depending on the disk amount you need.

I would make some benchmarks...
hth,
Philippe

We have had excellent stability on both G4 and G5, MacOS 10.2.x and
10.3.x,
PostgreSQL 7.3.x and 7.4.x. The only time we experienced instability
was just
after the G5 was released, the combination of G5, MacOS 10.2.7 and
PostgreSQL
7.3.x just didn't work very well. Upgrading the G5 to MacOS 10.3.x
and
PostgreSQL 7.4.x brought back the stability we expected and we
haven't really
had any problems since.

Our primary OS X 10.2 server crashed very frequently. Sometimes more
than once per day. We changed machines and the crashes continued.
Apple HW test on both boxes showed no problems. The vast majority of
these crashes were under moderate load (~120 queries/min). A few
times, reindexing would cause a crash without any other DB activity.
With almost all of these crashes, there were no CrashReporter log
entries. At that point, we felt like we had no recourse but to try
something different (Linux/x86) and haven't looked back.

- Jeff

--

Jeff Bohmer
VisionLink, Inc.
_______________ _______________ ___
303.402.0170 x121
http://www.visionlink.org/
_______________ _______________ ___
People. Tools. Change. Community.

---------------------------(end of
broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

[William Yu]

My guess is that you will get better performance from a similarly priced
Dual Opteron for the following reasons:

1) OS-X is not 64-bit yet, 64-bit Linux/BSD OS's are available
2) GCC is far better tuned for x86 than PowerPC/Itanium/etc
3) Postgres *seems* to prefer Opteron's ondie memory controller
architecture over shared bus -- especially in SMP configs

Is it enough of a difference? If you needed to eek out every possible %
performance because you have a critical production need, then the answer
is yes. Otherwise, it's easier to stick with the OS you know.
Jim Strickland wrote:

Well, the whole reason I have asked this question is because my
developer swears by OS X and PostgreSQL. However, I wanted opinions from
other people who have possibly used a similar setup so I can make an
informed decision. I will certainly keep your advice in mind. I guess
the only reason I was asking about the version of OS X and the G5
processor, is because that is all my developer uses and he seems to
think they make a great combination, but that seems to be at odds with
your experience.

Perhaps some others will weigh in with their experiences and I will be
able to make a sound decision. Fortunately there is no great rush to
decide. Thanks for your help.

[Timothy Perrigo]

I've been working with PostgreSQL on OS X (G4, G5 and dual G5 systems)
for a few months now, and overall I've been really pleased; for us, it
seems to be a good match.

If you have both an OS X and a Linux or BSD system available, you could
run pgbench against both and get a rough idea on how they compare.
On Nov 4, 2004, at 1:33 PM, William Yu wrote:

My guess is that you will get better performance from a similarly
priced Dual Opteron for the following reasons:

1) OS-X is not 64-bit yet, 64-bit Linux/BSD OS's are available
2) GCC is far better tuned for x86 than PowerPC/Itanium/etc
3) Postgres *seems* to prefer Opteron's ondie memory controller
architecture over shared bus -- especially in SMP configs

Is it enough of a difference? If you needed to eek out every possible
% performance because you have a critical production need, then the
answer is yes. Otherwise, it's easier to stick with the OS you know.
Jim Strickland wrote:

Well, the whole reason I have asked this question is because my
developer swears by OS X and PostgreSQL. However, I wanted opinions
from other people who have possibly used a similar setup so I can
make an informed decision. I will certainly keep your advice in mind.
I guess the only reason I was asking about the version of OS X and
the G5 processor, is because that is all my developer uses and he
seems to think they make a great combination, but that seems to be at
odds with your experience.
Perhaps some others will weigh in with their experiences and I will
be able to make a sound decision. Fortunately there is no great rush
to decide. Thanks for your help.

---------------------------(end of
broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to
ma*******@postg resql.org

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html