Re: [PERFORM] Tests
What I thought was really interesting in this test was a dramatic difference for ext3 mounted "writeback" in the "h1" test, 1 minute vs. 9 minutes compared to the default "ordered" mount option. This was the "add new constraint" test. --- Balazs Wellisch wrote: > > Nicely done! > > Thanks, > > Balazs > > > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Tomka > Gergely > Sent: Wednesday, August 27, 2003 5:40 AM > To: [EMAIL PROTECTED] > Subject: Re: [PERFORM] Tests > > http://mail.sth.sze.hu/~hsz/sql/ > > New, upgraded test results. As we see, the developers works hard, and > with > good results. Million thanks and congratulations. > > Sorry *BSD-lovers, if you send a new hard drive, our tester can do bsd > tests also. Life is hard. > > And last, but not least, thanks for the tests, Horv?th Szabolcs > > -- > Tomka Gergely > "S most - vajon barb?rok n?lk?l mi lesz vel?nk? > ?k m?giscsak megold?s voltak valahogy..." > > > ---(end of broadcast)--- > TIP 7: don't forget to increase your free space map settings > > > > ---(end of broadcast)--- > TIP 5: Have you checked our extensive FAQ? > >http://www.postgresql.org/docs/faqs/FAQ.html > -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [PERFORM] Tests
Nicely done! Thanks, Balazs -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Tomka Gergely Sent: Wednesday, August 27, 2003 5:40 AM To: [EMAIL PROTECTED] Subject: Re: [PERFORM] Tests http://mail.sth.sze.hu/~hsz/sql/ New, upgraded test results. As we see, the developers works hard, and with good results. Million thanks and congratulations. Sorry *BSD-lovers, if you send a new hard drive, our tester can do bsd tests also. Life is hard. And last, but not least, thanks for the tests, Horváth Szabolcs -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [PERFORM] Tests
http://mail.sth.sze.hu/~hsz/sql/ New, upgraded test results. As we see, the developers works hard, and with good results. Million thanks and congratulations. Sorry *BSD-lovers, if you send a new hard drive, our tester can do bsd tests also. Life is hard. And last, but not least, thanks for the tests, Horváth Szabolcs -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
Re: [PERFORM] Tests
2003-08-25 ragyogó napján scott.marlowe ezt üzente: > On Fri, 22 Aug 2003, Josh Berkus wrote: > > > Also another test I'd really like to see is hardware raid (Adaptec, LSI) > > against Linux software raid, and 5-disk RAID 5 against 4-disk RAID 1+0. > > It would be nice to cross those, so you have RAID5 sw vs RAID5 hw, vs > RAID1+0 sw vs RAID1+0 hw. > > Now if only I had the hardware. Time to scrounge. Hm. At first, now we dont have raid hardware, sorry. Second, in my practice the harware raid slower than sw raid, because now the cheapest computing power is the cpu. So i dont like hardware raid solutions, and even on ibm pc-servers i try not using them. -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [PERFORM] Tests
scott.marlowe wrote: On Fri, 22 Aug 2003, Josh Berkus wrote: Also another test I'd really like to see is hardware raid (Adaptec, LSI) against Linux software raid, and 5-disk RAID 5 against 4-disk RAID 1+0. It would be nice to cross those, so you have RAID5 sw vs RAID5 hw, vs RAID1+0 sw vs RAID1+0 hw. Now if only I had the hardware. Time to scrounge. I just wanted to comment, that I don't know where you guys find the time and money to do all this. I pretty much blew my entire weekend running tests, and I'm still not done enough to post any results. The upshot being, thanks to everyone who donates time, money, spare parts or whatever. Sometimes I forget how valuable all that is. And I don't think it gets said often enough how much people like me appreciate all the hard work the developers and testers and everyone else involved does. I mean, PostgreSQL (like Linux or BSD or any other open-source project) is a fantastic piece of software! -- Bill Moran Potential Technologies http://www.potentialtech.com ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] Tests
On Fri, 22 Aug 2003, Josh Berkus wrote: > Also another test I'd really like to see is hardware raid (Adaptec, LSI) > against Linux software raid, and 5-disk RAID 5 against 4-disk RAID 1+0. It would be nice to cross those, so you have RAID5 sw vs RAID5 hw, vs RAID1+0 sw vs RAID1+0 hw. Now if only I had the hardware. Time to scrounge. ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [PERFORM] Tests
2003-08-22 ragyogó napján Josh Berkus ezt üzente: > Tomka, > > Did you get my test suggestion? It never hit the lists, so I wonder if you > got it Yes, we got it, and we use your ideas. Thanks. -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] Tests
Tomka, Since this didn't post last time: Another good test to do would be one that measures simultaenous steaming read-write. For example: create table_a, with at least 10,000 records, where field1 is the PK and field2 is a text field. create table_b, with 500,000 or more records, where field2 is an FK to table_a, and field_3 is a text field. do the following: UPDATE table_b SET field3 = field3 || 'something' WHERE EXISTS (select field1 FROM table_a where table_a.field1 = table_b.field2 AND table_a.field_2 = 'x') For the test to be most effective, the condition above should affect about 100,000 records of table_a, and the records should not be contiguos on disk. Further, tablea.field2 should not be indexed in order to force a disk scan. Also another test I'd really like to see is hardware raid (Adaptec, LSI) against Linux software raid, and 5-disk RAID 5 against 4-disk RAID 1+0. -- -Josh Berkus Aglio Database Solutions San Francisco ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [PERFORM] Tests
Paul Thomas wrote: On 22/08/2003 22:00 Bill Moran wrote: [snip] Well, as much as I _am_ a die-hard BSD geek, I'm far more interested in knowing what platform is really best when I need a top-notch PostgreSQL server. I'm going to try to force some time this weekend to do some tests ... we'll see if I succeed ... I, for one, would be very interested to see your results (can't you find something better to at the weekend than about with computers ?:) Unfortunately, no. I've been trying to get a date for a month and I'm about ready to give up ;) What I'd really be interested in is a comparison of Linux vs BSD using each OS's variations of file system on the same single-processor Intel/AMD based hardware. That was what I was hoping to do. Selfishness and sillyness aside, I'm sure your tests will of interest to us all. Thanks in advance Oh great, now I feel obligated ... well, I guess I'd better get started. I'll post the results for as far as I get late sometime Sunday. I can't promise they'll be complete at that point, but we'll see what happens. -- Bill Moran Potential Technologies http://www.potentialtech.com ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [PERFORM] Tests
On 22/08/2003 22:00 Bill Moran wrote: [snip] Well, as much as I _am_ a die-hard BSD geek, I'm far more interested in knowing what platform is really best when I need a top-notch PostgreSQL server. I'm going to try to force some time this weekend to do some tests ... we'll see if I succeed ... I, for one, would be very interested to see your results (can't you find something better to at the weekend than about with computers ?:) What I'd really be interested in is a comparison of Linux vs BSD using each OS's variations of file system on the same single-processor Intel/AMD based hardware. Selfishness and sillyness aside, I'm sure your tests will of interest to us all. Thanks in advance -- Paul Thomas +--+-+ | Thomas Micro Systems Limited | Software Solutions for the Smaller Business | | Computer Consultants | http://www.thomas-micro-systems-ltd.co.uk | +--+-+ ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [PERFORM] Tests
On Fri, 2003-08-22 at 16:54, Tomka Gergely wrote:
> 2003-08-22 ragyogó napján Robert Treat ezt üzente:
>
> > On Thu, 2003-08-21 at 14:16, Bill Moran wrote:
> > > >>>What test are interesting? Plese give us tips and ideas. The guy has time
> > > >>>for other test.
> > > >>
> > > >>It's a shame you didn't test ufs+softupdates. I'd be curious to see how
> > > >>it stacked up against the others.
> > > >
> > > > Shame? I smell here a harcore BSD fighter :)
> > >
> > > Well, I suppose ...
> > >
> >
> > Well I'm not a hardcore bsd fighter and I'd like to see how it stacks up
> > as well. UFS+softupdates is supposed to be a very safe combination, if
> > it performs well enough I could see a recommendation for it for those
> > who are willing to look beyond linux.
>
> The guy who do the test have only a few weeks *bsd-experience, and i dont
> have bsd experience at all. The guy now planning tests on BSD, but he
> need some time to build up a good relationship with the *BSD.
>
Another thought would be linux w/ xfs
Also, can you post more complete hardware/ os info?
Oh, and vs 7.4beta1 would be great too.
:-)
Robert Treat
--
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
Re: [PERFORM] Tests
Tomka, Did you get my test suggestion? It never hit the lists, so I wonder if you got it -- -Josh Berkus Aglio Database Solutions San Francisco ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [PERFORM] Tests
Tomka Gergely wrote: 2003-08-22 ragyogó napján Robert Treat ezt üzente: On Thu, 2003-08-21 at 14:16, Bill Moran wrote: What test are interesting? Plese give us tips and ideas. The guy has time for other test. It's a shame you didn't test ufs+softupdates. I'd be curious to see how it stacked up against the others. Shame? I smell here a harcore BSD fighter :) Well, I suppose ... Well I'm not a hardcore bsd fighter and I'd like to see how it stacks up as well. UFS+softupdates is supposed to be a very safe combination, if it performs well enough I could see a recommendation for it for those who are willing to look beyond linux. The guy who do the test have only a few weeks *bsd-experience, and i dont have bsd experience at all. The guy now planning tests on BSD, but he need some time to build up a good relationship with the *BSD. Ps.: i am a harcore linux-fgihter, so sorry :) Well, as much as I _am_ a die-hard BSD geek, I'm far more interested in knowing what platform is really best when I need a top-notch PostgreSQL server. I'm going to try to force some time this weekend to do some tests ... we'll see if I succeed ... -- Bill Moran Potential Technologies http://www.potentialtech.com ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] Tests
2003-08-22 ragyogó napján Robert Treat ezt üzente: > On Thu, 2003-08-21 at 14:16, Bill Moran wrote: > > >>>What test are interesting? Plese give us tips and ideas. The guy has time > > >>>for other test. > > >> > > >>It's a shame you didn't test ufs+softupdates. I'd be curious to see how > > >>it stacked up against the others. > > > > > > Shame? I smell here a harcore BSD fighter :) > > > > Well, I suppose ... > > > > Well I'm not a hardcore bsd fighter and I'd like to see how it stacks up > as well. UFS+softupdates is supposed to be a very safe combination, if > it performs well enough I could see a recommendation for it for those > who are willing to look beyond linux. The guy who do the test have only a few weeks *bsd-experience, and i dont have bsd experience at all. The guy now planning tests on BSD, but he need some time to build up a good relationship with the *BSD. Ps.: i am a harcore linux-fgihter, so sorry :) -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [PERFORM] Tests
On Thu, 2003-08-21 at 14:16, Bill Moran wrote:
> >>>What test are interesting? Plese give us tips and ideas. The guy has time
> >>>for other test.
> >>
> >>It's a shame you didn't test ufs+softupdates. I'd be curious to see how
> >>it stacked up against the others.
> >
> > Shame? I smell here a harcore BSD fighter :)
>
> Well, I suppose ...
>
Well I'm not a hardcore bsd fighter and I'd like to see how it stacks up
as well. UFS+softupdates is supposed to be a very safe combination, if
it performs well enough I could see a recommendation for it for those
who are willing to look beyond linux.
Robert Treat
--
Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [PERFORM] Tests
Tomka Gergely wrote: 2003-08-21 ragyogó napján Bill Moran ezt üzente: Tomka Gergely wrote: http://195.199.65.92/~horvaths/pgperformance.html One of my friend do these tests. We think about the best filesystem for the Linux/Postgres systems. The four test was: teszt1 - 10.000 inserts teszt2 - 10.inserts, 10 in one trans. teszt3 - 14.000.000 inserts, 1.000 in one trans. teszt4 - "Select count(*)" on the big table. What test are interesting? Plese give us tips and ideas. The guy has time for other test. It's a shame you didn't test ufs+softupdates. I'd be curious to see how it stacked up against the others. Shame? I smell here a harcore BSD fighter :) Well, I suppose ... Maybe sometime he run tests on *SD, but now we need linux data. "He" has had it on his list of things to do for months. And if I ever get enough free moments, I will test *both* BSD and Linux, because I'm terribly biased towards BSD because that's all I use. In addition to being interested in how FreeBSD's ufs+softupdates compares to the variety of Linux journalling filesystems, I'm also interested to compare FreeBSD's performance on ext2fs and FAT to Linux's peformance on those same filesystems. It's a "shame" because I simply haven't found the time yet. -- Bill Moran Potential Technologies http://www.potentialtech.com ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [PERFORM] Tests
2003-08-21 ragyogó napján Bill Moran ezt üzente: > Tomka Gergely wrote: > > http://195.199.65.92/~horvaths/pgperformance.html > > > > One of my friend do these tests. We think about the best filesystem for > > the Linux/Postgres systems. > > > > The four test was: > > > > teszt1 - 10.000 inserts > > teszt2 - 10.inserts, 10 in one trans. > > teszt3 - 14.000.000 inserts, 1.000 in one trans. > > teszt4 - "Select count(*)" on the big table. > > > > What test are interesting? Plese give us tips and ideas. The guy has time > > for other test. > > It's a shame you didn't test ufs+softupdates. I'd be curious to see how > it stacked up against the others. Shame? I smell here a harcore BSD fighter :) Maybe sometime he run tests on *SD, but now we need linux data. -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [PERFORM] Tests
Tomka Gergely wrote: http://195.199.65.92/~horvaths/pgperformance.html One of my friend do these tests. We think about the best filesystem for the Linux/Postgres systems. The four test was: teszt1 - 10.000 inserts teszt2 - 10.inserts, 10 in one trans. teszt3 - 14.000.000 inserts, 1.000 in one trans. teszt4 - "Select count(*)" on the big table. What test are interesting? Plese give us tips and ideas. The guy has time for other test. It's a shame you didn't test ufs+softupdates. I'd be curious to see how it stacked up against the others. -- Bill Moran Potential Technologies http://www.potentialtech.com ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [PERFORM] Tests
On 21 Aug 2003 at 16:51, Tomka Gergely wrote: > http://195.199.65.92/~horvaths/pgperformance.html > > One of my friend do these tests. We think about the best filesystem for > the Linux/Postgres systems. > > The four test was: > > teszt1 - 10.000 inserts > teszt2 - 10.inserts, 10 in one trans. > teszt3 - 14.000.000 inserts, 1.000 in one trans. > teszt4 - "Select count(*)" on the big table. > > What test are interesting? Plese give us tips and ideas. The guy has time > for other test. - Try creating table without OIDs and repeat runs - Try changing number of inserts per transactions. I tried similar test on my home machine last week. It gave me 500K records in about 54 seconds on 7.3.3. That translates to roughly 3 minutes for 1.4M records. You sure something isn't right here? What is the hardware? Can you try it with 7.4CVS? Bye Shridhar -- XIIdigitation, n.: The practice of trying to determine the year a movie was madeby deciphering the Roman numerals at the end of the credits.-- Rich Hall, "Sniglets" ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
[PERFORM] Tests
http://195.199.65.92/~horvaths/pgperformance.html One of my friend do these tests. We think about the best filesystem for the Linux/Postgres systems. The four test was: teszt1 - 10.000 inserts teszt2 - 10.inserts, 10 in one trans. teszt3 - 14.000.000 inserts, 1.000 in one trans. teszt4 - "Select count(*)" on the big table. What test are interesting? Plese give us tips and ideas. The guy has time for other test. -- Tomka Gergely "S most - vajon barbárok nélkül mi lesz velünk? Ők mégiscsak megoldás voltak valahogy..." ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
