Thread: What is the recommended machine configuration?
We are planning to dedicate a new Pentium II/400 MHz with Red Hat Linux exclusively to our Postgres database. The database size is several hundred megs. I have been told that the more RAM I have for the machine, the faster I'll be able to access the data. Does anyone have any recommendations on the amount of RAM that would be optimal? If you had to set up the ideal system to run a Postgres database of, let's say, 1 Gig in size (spread across 10 tables), how would you allocate it in terms of processor, OS, RAM, hard disk space, etc.? Thanks. -Tony Reina
> > We are planning to dedicate a new Pentium II/400 MHz with Red Hat Linux > exclusively to our Postgres database. The database size is several > hundred megs. I have been told that the more RAM I have for the machine, > > the faster I'll be able to access the data. Does anyone have any > recommendations on the amount of RAM that would be optimal? If you had > to set up the ideal system to run a Postgres database of, let's say, 1 > Gig in size (spread across 10 tables), how would you allocate it in > terms of processor, OS, RAM, hard disk space, etc.? > I'm not much of an expert but the system I'm putting together has between 15-20 tables, comprising about 250k rows totaling about 1.5-2GB of disk space. Right now I'm running under BSDi 3.0 with 128MB RAM on a PPRO 200. Performance is great...almost no issues.... Pierre
On Wed, 5 Aug 1998, G. Anthony Reina wrote: > We are planning to dedicate a new Pentium II/400 MHz with Red Hat Linux > exclusively to our Postgres database. The database size is several > hundred megs. I have been told that the more RAM I have for the machine, > the faster I'll be able to access the data. Does anyone have any > recommendations on the amount of RAM that would be optimal? If you had > to set up the ideal system to run a Postgres database of, let's say, 1 > Gig in size (spread across 10 tables), how would you allocate it in > terms of processor, OS, RAM, hard disk space, etc.? I'm going to poke my head up here, as carefully as I can. This is *not* meant to start a flame war, I'm just bringing up a point that someone else made a short while ago.. There was a discussion a little while back (probably prompted by me) concerning Linux vs FreeBSD, and their "intended uses". One of the comments made by a *Linux* user went to the effect that Linux makes a great desktop computer (ie. MicroSloth Windows replacement), but if you want a *server* operating system, go with FreeBSD, as it scales better. This was a comment from a Linux user...there are several Linux users on the lists, and Thomas (one of the core developers) is one of them, so it isn't as if PostgreSQL isn't *extensively* tested under Linux... One thing in RedHat's favor, though, is that PostgreSQL comes with the operating system as an installable package... Marc G. Fournier Systems Administrator @ hub.org primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
One thing in RedHat's favor, though, is that PostgreSQL comes with the operating system as an installable package... So does FreeBSD, doesn't it? I know NetBSD does. Cheers, Brook
On Wed, 5 Aug 1998, Brook Milligan wrote: > One thing in RedHat's favor, though, is that PostgreSQL comes with the > operating system as an installable package... > > So does FreeBSD, doesn't it? I know NetBSD does. We have it as part of "ports", which means that as long as you haven't done anything weird to your machine, it will compile, but compile it you have to do. RedHat has it as a pre-compiled package that you just install just as you would any other package, when you install the operating system... Marc G. Fournier Systems Administrator @ hub.org primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
Re: [SQL] What is the recommended machine configuration?
From
darcy@druid.net (D'Arcy J.M. Cain)
Date:
Thus spake The Hermit Hacker > One thing in RedHat's favor, though, is that PostgreSQL comes with the > operating system as an installable package... With NetBSD too. I would be surprised if it wasn't in the FreeBSD packages as well. -- D'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves http://www.druid.net/darcy/ | and a sheep voting on +1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.
Hi all, I like this mailing list, very helpful. I have a question regarding splitting data file to more than 1 hardisks. Currently I have 2 x 2 GB hardisks. I would like to split files in directory /usr/local/pgsql/data/base/databasename to 2 hardisks. The reason is I want to dump 3 GB ~ 4 GB of data from oracle to postgres to do testing. I want to know whether postgres can handle 3 GB ~ 4 GB data, and so some performance benchmarking. ( Oracle vs Postgres ) I don't have one 4 GB hardisk, so I have to split them. I am using Linux. Could any one help me ? Thanks Regards Chai
On Thu, 6 Aug 1998, Chairudin Sentosa Harjo wrote: > Hi all, > > I like this mailing list, very helpful. > I have a question regarding splitting data file to more than 1 hardisks. > > Currently I have 2 x 2 GB hardisks. > I would like to split files in directory > /usr/local/pgsql/data/base/databasename > to 2 hardisks. > > The reason is I want to dump 3 GB ~ 4 GB of data from oracle to postgres to > do testing. > I want to know whether postgres can handle 3 GB ~ 4 GB data, and so some > performance benchmarking. ( Oracle vs Postgres ) > > I don't have one 4 GB hardisk, so I have to split them. > I am using Linux. > Could any one help me ? > > Thanks > > Regards > Chai > Hi, I don't know if PostgreSQL can do that.. but you could certainly enabled the linear mode for 'disk appending' in your kernel. Hope that helps, Ludovic
Ludovic Marcotte wrote: > > On Thu, 6 Aug 1998, Chairudin Sentosa Harjo wrote: > > > Hi all, > > > > I like this mailing list, very helpful. > > I have a question regarding splitting data file to more than 1 hardisks. > > > > Currently I have 2 x 2 GB hardisks. > > I would like to split files in directory > > /usr/local/pgsql/data/base/databasename > > to 2 hardisks. > > > > The reason is I want to dump 3 GB ~ 4 GB of data from oracle to postgres to > > do testing. > > I want to know whether postgres can handle 3 GB ~ 4 GB data, and so some > > performance benchmarking. ( Oracle vs Postgres ) > > > > I don't have one 4 GB hardisk, so I have to split them. > > I am using Linux. > > Could any one help me ? > > > > Thanks > > > > Regards > > Chai > > > > Hi, > I don't know if PostgreSQL can do that.. but you could certainly > enabled the linear mode for 'disk appending' in your kernel. or use software RAID (a.k.a. "disk striping") wich may nealy double your disk performance if you use SCSI or attach IDE drives to separate controllers. http://linas.org/linux/Software-RAID/Software-RAID.html Gene > > Hope that helps, > Ludovic
Hi all, I started the postmaster by running this command: /usr/local/psql/bin/postmaster -oe -S -D /usr/local/pgsql/data Then I type: psql mydb Error: Connection to database 'mydb' failed. PQexec() -- Request was sent to backend, but backed closed the channel before responding. This probably means the backend terminated abnormally before or while processing request. What is the proper way to terminate the postmaster? Regards Chai
On Wed, 5 Aug 1998, D'Arcy J.M. Cain wrote: > Thus spake The Hermit Hacker > > One thing in RedHat's favor, though, is that PostgreSQL comes with the > > operating system as an installable package... > > With NetBSD too. I would be surprised if it wasn't in the FreeBSD > packages as well. And we are working on this for a future release of slackware... Cheers, Tom ----------- Sisters of Charity Medical Center ---------- Department of Psychiatry ---- Thomas Good, System Administrator <tomg@q8.nrnet.org> North Richmond CMHC/Residential Services Phone: 718-354-5528 75 Vanderbilt Ave, Quarters 8 Fax: 718-354-5056 Staten Island, NY 10304 www.panix.com/~ugd ---- Powered by PostgreSQL 6.3.2 / Perl 5.004 / DBI-0.91::DBD-PG-0.69
I remember reading some part of the docs I that mentioned a significant difference in speed of queries between 64 bit and 32 bit architectures. If you are planning on using Red Hat Linux you could just as easily swap the PII for an Alpha and get better performance, higher IO throughput and an all around better machine. FYI - RedHat ships with the following 4 postgres related packages on the alpha: postgresql-6.3.2-4.alpha.rpm postgresql-clients-6.3.2-4.alpha.rpm postgresql-data-6.3.2-4.alpha.rpm postgresql-devel-6.3.2-4.alpha.rpm Anyone with experience using RedHat on Alpha with Postgres care to comment? Chris On Wed, 5 Aug 1998, G. Anthony Reina wrote: > We are planning to dedicate a new Pentium II/400 MHz with Red Hat Linux > exclusively to our Postgres database. The database size is several > hundred megs. I have been told that the more RAM I have for the machine, > > the faster I'll be able to access the data. Does anyone have any > recommendations on the amount of RAM that would be optimal? If you had > to set up the ideal system to run a Postgres database of, let's say, 1 > Gig in size (spread across 10 tables), how would you allocate it in > terms of processor, OS, RAM, hard disk space, etc.? > > Thanks. > -Tony Reina > >
On Thu, 6 Aug 1998, Chairudin Sentosa Harjo wrote: > Hi all, > > I like this mailing list, very helpful. > I have a question regarding splitting data file to more than 1 hardisks. > > Currently I have 2 x 2 GB hardisks. > I would like to split files in directory > /usr/local/pgsql/data/base/databasename > to 2 hardisks. > > The reason is I want to dump 3 GB ~ 4 GB of data from oracle to postgres to > do testing. > I want to know whether postgres can handle 3 GB ~ 4 GB data, and so some > performance benchmarking. ( Oracle vs Postgres ) > > I don't have one 4 GB hardisk, so I have to split them. > I am using Linux. > Could any one help me ? If your using Linux you can throw those 2 disks in a RAID configuration as to form one logical disk. You can try striping (RAID 1???) or RAID5 I guess.... Maarten _____________________________________________________________________________ | TU Delft, The Netherlands, Faculty of Information Technology and Systems | | Department of Electrical Engineering | | Computer Architecture and Digital Technique section | | M.Boekhold@et.tudelft.nl | -----------------------------------------------------------------------------
> Hi all, > > I started the postmaster by running this command: > /usr/local/psql/bin/postmaster -oe -S -D /usr/local/pgsql/data > > Then I type: > psql mydb > > Error: > Connection to database 'mydb' failed. > PQexec() -- Request was sent to backend, but backed closed the channel before > responding. > This probably means the backend terminated abnormally > before or while > processing request. > What user are you logged in as when you try to connect to the database? If you are not a users that has also had a postgres account created by create_user this may happen. Also, this or other error messages will occur if you try to access the database as root...james
On Wed, 5 Aug 1998, The Hermit Hacker wrote: > One thing in RedHat's favor, though, is that PostgreSQL comes with the > operating system as an installable package... That is also true of Debian Linux. Johann -------------------------------------------------------------------------- | Johann Spies Windsorlaan 19 | | jhspies@alpha.futurenet.co.za 3201 Pietermaritzburg | | Tel/Faks Nr. +27 331-46-1310 Suid-Afrika (South Africa) | -------------------------------------------------------------------------- "Do not be anxious about anything; but in every thing by prayer and supplication with thanksgiving let your requests be made known unto God." Philippians 4:6