Thread: Location of PostgreSQL Performance Test Lab
Folks, In addition to the machine Gavin Roy is loaning, I'm putting together an online performance test lab for the commmunity using donated equipment. Fairly soon, however, I have to decide where it will be hosted. Since the plurality of the equipment is coming from California, I'd prefer something on the West Coast, but the critical issue is somewhere where we have ready access to sysadmin help to connect things, replace drives etc. I was thinking of OSL in Corvalis Oregon, for a few reasons: -- Easy shipping -- Full-time sysadmin staff -- One machine is already there -- Close to Mark Wong -- Well-known for OSS hosting However, OSL has the drawback that we would need to raise money to cover power costs for the new machines, probably a couple $thousand. Other locations which we have here in the US include: -- ArcServe in San Jose (Sean Chittenden) -- EHPG in Fremont (Gavin Roy) -- Someone in Seattle? Note that I'm talking about between 1/2 and a full rack of equipment here ... not just a couple servers. I'd prefer not to have the servers scattered hither and yon, either, if I can avoid it. -- Josh Berkus PostgreSQL @ Sun San Francisco
Am Dienstag, 31. Juli 2007 18:19 schrieb Josh Berkus: > In addition to the machine Gavin Roy is loaning, I'm putting together an > online performance test lab for the commmunity using donated equipment. > Fairly soon, however, I have to decide where it will be hosted. Why is it important that all the machines be hosted at one site? -- Peter Eisentraut http://developer.postgresql.org/~petere/
Peter, > > In addition to the machine Gavin Roy is loaning, I'm putting together an > > online performance test lab for the commmunity using donated equipment. > > Fairly soon, however, I have to decide where it will be hosted. > > Why is it important that all the machines be hosted at one site? Administration and the ability to to do multi-machine configurations. Also the ability to swap storage to newer machines. -- Josh Berkus PostgreSQL @ Sun San Francisco
Josh Berkus wrote: > Folks, > > In addition to the machine Gavin Roy is loaning, I'm putting together an > online performance test lab for the commmunity using donated equipment. > Fairly soon, however, I have to decide where it will be hosted. Since the > plurality of the equipment is coming from California, I'd prefer something on > the West Coast, but the critical issue is somewhere where we have ready > access to sysadmin help to connect things, replace drives etc. > > I was thinking of OSL in Corvalis Oregon, for a few reasons: > -- Easy shipping > -- Full-time sysadmin staff > -- One machine is already there > -- Close to Mark Wong > -- Well-known for OSS hosting > However, OSL has the drawback that we would need to raise money to cover power > costs for the new machines, probably a couple $thousand. per year I assume ? Stefan
Am Dienstag, 31. Juli 2007 18:43 schrieb Josh Berkus: > Peter, > > > > In addition to the machine Gavin Roy is loaning, I'm putting together > > > an online performance test lab for the commmunity using donated > > > equipment. Fairly soon, however, I have to decide where it will be > > > hosted. > > > > Why is it important that all the machines be hosted at one site? > > Administration Why does administration require that all the machines be hosted at one site? Most administration is by remote login, and for the physical maintenance you could have someone on site do it. > and the ability to to do multi-machine configurations. Well, that is interesting, but not a requirement. If the machines are not at the same site, just don't use them for multi-machine configurations; just pick different ones. (Hey, maybe someone wants to test WAN configurations.) > Also the ability to swap storage to newer machines. I don't know what you mean by this. -- Peter Eisentraut http://developer.postgresql.org/~petere/
Peter, > Why does administration require that all the machines be hosted at one > site? Most administration is by remote login, and for the physical > maintenance you could have someone on site do it. Right, and that's made more complex if the main users aren't clear on where the machine is. > > and the ability to to do multi-machine configurations. > > Well, that is interesting, but not a requirement. If the machines are not > at the same site, just don't use them for multi-machine configurations; > just pick different ones. (Hey, maybe someone wants to test WAN > configurations.) Actually, it *is* a requirement. Doing a real DBT5 or EAStress test run requires at least 2 machines, plus storage, and 3-4 machines is more interesting. Having the machines scattered hither and yon will both complicate the maintenance of the test lab and limit its usefulness. I also think you're arguing for the sake of arguing; you don't do performance testing and AFAIK are not planning to host any of the machines. So why are you picking nits? -- Josh Berkus PostgreSQL @ Sun San Francisco
Josh Berkus wrote: > Right, and that's made more complex if the main users aren't clear on > where the machine is. You have certainly heard of the world wide web, where you can easily publish fabulous lists of information in easily accessible form. > Actually, it *is* a requirement. Doing a real DBT5 or EAStress test > run requires at least 2 machines, plus storage, and 3-4 machines is > more interesting. Having the machines scattered hither and yon will > both complicate the maintenance of the test lab and limit its > usefulness. A requirement to have 4 machines at one site to perform a particular test is not a logical reason to reject the possibility of having, say, 20 machines at 5 sites to perform all kinds of tests, when the alternative is to have no machines at no sites. If you can find someone to host everything, great, but you have in fact already rejected two machines because they are not at the right location, when I can see great use for those machines no matter where they are. > I also think you're arguing for the sake of arguing; you don't do > performance testing and AFAIK are not planning to host any of the > machines. I think you are wrong. -- Peter Eisentraut http://developer.postgresql.org/~petere/
Peter, > A requirement to have 4 machines at one site to perform a particular > test is not a logical reason to reject the possibility of having, say, > 20 machines at 5 sites to perform all kinds of tests, when the > alternative is to have no machines at no sites. Ah, I see the misunderstanding. Would that we were talking about 20 machines. Right now I'm twisting arms to get 5 servers and 4 JBODs, which we need in the same location to make use of them. So we need these initial machines in one place, as much as possible, to make use of them. > If you can find someone to host everything, great, but you have in fact > already rejected two machines because they are not at the right > location, when I can see great use for those machines no matter where > they are. The peformance testing we can do without storage (drives) is very limited. Currently, the only storage I have is located in the US. In my experience, storage is the hardest thing to get donated so I don't really expect to find some in Europe soon. Gavin's machine is the exception becuase it comes with a 28-disk array. Still, we could do even more with it if we had another machine or two in that location. -- --Josh Josh Berkus PostgreSQL @ Sun San Francisco
That could possibly be arranged, let me make some calls... 1/2 cab? On 7/31/07, Josh Berkus <josh@agliodbs.com> wrote: > Peter, > > > A requirement to have 4 machines at one site to perform a particular > > test is not a logical reason to reject the possibility of having, say, > > 20 machines at 5 sites to perform all kinds of tests, when the > > alternative is to have no machines at no sites. > > Ah, I see the misunderstanding. Would that we were talking about 20 > machines. Right now I'm twisting arms to get 5 servers and 4 JBODs, which > we need in the same location to make use of them. So we need these > initial machines in one place, as much as possible, to make use of them. > > > If you can find someone to host everything, great, but you have in fact > > already rejected two machines because they are not at the right > > location, when I can see great use for those machines no matter where > > they are. > > The peformance testing we can do without storage (drives) is very limited. > Currently, the only storage I have is located in the US. In my > experience, storage is the hardest thing to get donated so I don't really > expect to find some in Europe soon. > > Gavin's machine is the exception becuase it comes with a 28-disk array. > Still, we could do even more with it if we had another machine or two in > that location. > > -- > --Josh > > Josh Berkus > PostgreSQL @ Sun > San Francisco > > ---------------------------(end of broadcast)--------------------------- > TIP 3: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faq >
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 - --On Tuesday, July 31, 2007 20:00:38 +0200 Peter Eisentraut <peter_e@gmx.net> wrote: > Josh Berkus wrote: >> Right, and that's made more complex if the main users aren't clear on >> where the machine is. > > You have certainly heard of the world wide web, where you can easily > publish fabulous lists of information in easily accessible form. Have to agree with Peter on this one ... not necessarily with as much sarcasm, of course ... that one is a weak argument ... > If you can find someone to host everything, great, but you have in fact > already rejected two machines because they are not at the right > location, when I can see great use for those machines no matter where > they are. Have to agree here as well ... sounds like you are looking for utopia where we generally live in a world of limbo ... >> I also think you're arguing for the sake of arguing; you don't do >> performance testing and AFAIK are not planning to host any of the >> machines. > > I think you are wrong. And on this one, I definitely agree, but for different reasons ... there is nobody on this list that a) can provide hosting for this or b) you would host with ... and those that will/would do the performance testing don't care in the least if the servers are in California or Timbuktoo ... so if you didn't want discussion on this, why bother posting it here ... ? - ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . scrappy@hub.org MSN . scrappy@hub.org Yahoo . yscrappy Skype: hub.org ICQ . 7615664 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFGr4MX4QvfyHIvDvMRAl2HAJ9j6q6/1sOSunVlcnfp3v9RXkXpxACgs11/ jRZ3UmiROthE4/+0EjREw0k= =VbZm -----END PGP SIGNATURE-----
Peter Eisentraut wrote:
> Am Dienstag, 31. Juli 2007 18:43 schrieb Josh Berkus:
>> Peter,
>>
>>>> In addition to the machine Gavin Roy is loaning, I'm putting together
>>>> an online performance test lab for the commmunity using donated
>>>> equipment. Fairly soon, however, I have to decide where it will be
>>>> hosted.
>>> Why is it important that all the machines be hosted at one site?
>> Administration
>
> Why does administration require that all the machines be hosted at one site?
> Most administration is by remote login, and for the physical maintenance you
> could have someone on site do it.
>
>> and the ability to to do multi-machine configurations.
>
> Well, that is interesting, but not a requirement. If the machines are not at
> the same site, just don't use them for multi-machine configurations; just
> pick different ones. (Hey, maybe someone wants to test WAN configurations.)
>
>> Also the ability to swap storage to newer machines.
>
> I don't know what you mean by this.
Often IO has not improved, while the cage has (more ram, more
processors, better cache etc)... So we can move the 22 spindle array to
a new machine easily.
Joshua D. Drake
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
Gavin M. Roy wrote: > That could possibly be arranged, let me make some calls... 1/2 cab? It should be noted that: A. Command Prompt has physical access to its location B. Command Prompt has tons of bandwidth C. Command Prompt has power D. Command Prompt has proven a reliable provider E. Command Prompt loves the PostgreSQL community ;) F. We would probably have to charge a bit for power cause we are talking about some serious machines. G. We are close to Mark Wong as well Sincerely, Joshua D. Drake > > On 7/31/07, Josh Berkus <josh@agliodbs.com> wrote: >> Peter, >> >>> A requirement to have 4 machines at one site to perform a particular >>> test is not a logical reason to reject the possibility of having, say, >>> 20 machines at 5 sites to perform all kinds of tests, when the >>> alternative is to have no machines at no sites. >> Ah, I see the misunderstanding. Would that we were talking about 20 >> machines. Right now I'm twisting arms to get 5 servers and 4 JBODs, which >> we need in the same location to make use of them. So we need these >> initial machines in one place, as much as possible, to make use of them. >> >>> If you can find someone to host everything, great, but you have in fact >>> already rejected two machines because they are not at the right >>> location, when I can see great use for those machines no matter where >>> they are. >> The peformance testing we can do without storage (drives) is very limited. >> Currently, the only storage I have is located in the US. In my >> experience, storage is the hardest thing to get donated so I don't really >> expect to find some in Europe soon. >> >> Gavin's machine is the exception becuase it comes with a 28-disk array. >> Still, we could do even more with it if we had another machine or two in >> that location. >> >> -- >> --Josh >> >> Josh Berkus >> PostgreSQL @ Sun >> San Francisco >> >> ---------------------------(end of broadcast)--------------------------- >> TIP 3: Have you checked our extensive FAQ? >> >> http://www.postgresql.org/docs/faq >> > > ---------------------------(end of broadcast)--------------------------- > TIP 6: explain analyze is your friend > -- === The PostgreSQL Company: Command Prompt, Inc. === Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240 Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/ Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate PostgreSQL Replication: http://www.commandprompt.com/products/
> Note that I'm talking about between 1/2 and a full rack of equipment here ...
> not just a couple servers. I'd prefer not to have the servers scattered
> hither and yon, either, if I can avoid it.
See my post please.
Joshua D. Drake
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
JD, > A. Command Prompt has physical access to its location > B. Command Prompt has tons of bandwidth > C. Command Prompt has power > D. Command Prompt has proven a reliable provider > E. Command Prompt loves the PostgreSQL community ;) > F. We would probably have to charge a bit for power cause we are talking > about some serious machines. > G. We are close to Mark Wong as well OK, I'm sold, barring objections from the donors. Anyone? -- --Josh Josh Berkus PostgreSQL @ Sun San Francisco
Josh Berkus wrote: > JD, > >> A. Command Prompt has physical access to its location >> B. Command Prompt has tons of bandwidth >> C. Command Prompt has power >> D. Command Prompt has proven a reliable provider >> E. Command Prompt loves the PostgreSQL community ;) >> F. We would probably have to charge a bit for power cause we are talking >> about some serious machines. >> G. We are close to Mark Wong as well > > OK, I'm sold, barring objections from the donors. Anyone? > How much of Command Prompt have access to the location? We've been waiting for a number of weeks to get the new pgFoundry machine plugged in and switched on because only JD had access to the DC and he was busy preparing for OSCON. I don't object in principle, I'd just like some reassurance that we won't see the perf machines lying idle or awaiting maintenance for the same or similar reasons. Regards, Dave
On Wed, Aug 01, 2007 at 08:45:56AM +0100, Dave Page wrote: > Josh Berkus wrote: > > JD, > > > >> A. Command Prompt has physical access to its location > >> B. Command Prompt has tons of bandwidth > >> C. Command Prompt has power > >> D. Command Prompt has proven a reliable provider > >> E. Command Prompt loves the PostgreSQL community ;) > >> F. We would probably have to charge a bit for power cause we are talking > >> about some serious machines. > >> G. We are close to Mark Wong as well > > > > OK, I'm sold, barring objections from the donors. Anyone? > > > > How much of Command Prompt have access to the location? We've been > waiting for a number of weeks to get the new pgFoundry machine plugged > in and switched on because only JD had access to the DC and he was busy > preparing for OSCON. > > I don't object in principle, I'd just like some reassurance that we > won't see the perf machines lying idle or awaiting maintenance for the > same or similar reasons. Yeah, that's especially important given the talk about being able to easily move disk array between servers and such. While we don't need 30-minute-on-site committment, we do need some level of commitment for something like that. //Magnus
Dave, all: > I don't object in principle, I'd just like some reassurance that we > won't see the perf machines lying idle or awaiting maintenance for the > same or similar reasons. That is the advantage of OSL, which has on-site staff and are pretty much immediately available during US working hours. --Josh Berkus
Dave Page wrote:
> Josh Berkus wrote:
>> JD,
> How much of Command Prompt have access to the location? We've been
> waiting for a number of weeks to get the new pgFoundry machine plugged
> in and switched on because only JD had access to the DC and he was busy
> preparing for OSCON.
>
> I don't object in principle, I'd just like some reassurance that we
> won't see the perf machines lying idle or awaiting maintenance for the
> same or similar reasons.
Sure, but I have mentioned that this would be a bad time of year for the
pgfoundry machine. Basically I had prep for oscon, oscon, and then I am
teaching a class ;). This was the *worst* time of year to try and prep
something new.
But to answer the question, I have to provision the power and rack space
but other than that we have immediate access to the location that does
not require me.
Joshua D. Drake
>
> Regards, Dave
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
Magnus Hagander wrote:
> On Wed, Aug 01, 2007 at 08:45:56AM +0100, Dave Page wrote:
>> Josh Berkus wrote:
>>> JD,
>> I don't object in principle, I'd just like some reassurance that we
>> won't see the perf machines lying idle or awaiting maintenance for the
>> same or similar reasons.
>
> Yeah, that's especially important given the talk about being able to easily
> move disk array between servers and such. While we don't need
> 30-minute-on-site committment, we do need some level of commitment for
> something like that.
As said, it is only for provisioning that I am required.
Joshua D. Drake
>
> //Magnus
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: Don't 'kill -9' the postmaster
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
Joshua D. Drake wrote: > Sure, but I have mentioned that this would be a bad time of year for the > pgfoundry machine. Basically I had prep for oscon, oscon, and then I am > teaching a class ;). This was the *worst* time of year to try and prep > something new. > > But to answer the question, I have to provision the power and rack space > but other than that we have immediate access to the location that does > not require me. So just to be clear, if a couple of disks die and we need to reinstall the OS on new ones, there are people other than you who can replace the disks and perform a base install to get us back up and running again within a day or two? You personally are *only* required for the initial physical install into the rack? Regards, Dave
On Wed, Aug 01, 2007 at 09:44:21PM -0700, Joshua D. Drake wrote: > Dave Page wrote: > >Josh Berkus wrote: > >>JD, > > >How much of Command Prompt have access to the location? We've been > >waiting for a number of weeks to get the new pgFoundry machine plugged > >in and switched on because only JD had access to the DC and he was busy > >preparing for OSCON. > > > >I don't object in principle, I'd just like some reassurance that we > >won't see the perf machines lying idle or awaiting maintenance for the > >same or similar reasons. > > Sure, but I have mentioned that this would be a bad time of year for the > pgfoundry machine. Basically I had prep for oscon, oscon, and then I am > teaching a class ;). This was the *worst* time of year to try and prep > something new. > > But to answer the question, I have to provision the power and rack space > but other than that we have immediate access to the location that does > not require me. Good - that makes it a much nicer option ;) But to bug you a bit further - how many other people ware we talking about, within reasonable distance? Or does the datacenter have on-call people locally? (I mean, it's not like we can call on Devrim to go there and plug something in :-P) //Magnus
Magnus Hagander wrote: > Good - that makes it a much nicer option ;) > > But to bug you a bit further - how many other people ware we talking about, > within reasonable distance? Or does the datacenter have on-call people > locally? We have two people that are CMD people that are local to the machines. However the data center has people in house (not just on call but 20 feet away) and any thing that needs to be done 8-6PST could be done via phone. It is actually how CMD does *everything* now with the exception of provisioning which I reserve for myself to make sure everything stays sane. > > (I mean, it's not like we can call on Devrim to go there and plug something > in :-P) Nope ;) but we can request a plug in or swap or whatever. When the machines show up we will label and document everything so it will be basically legos for the data center folks. Oh and just so everyone understands the relationship with the data center, I started it... years before CMD existed. I drink beer and golf with the owners, I am invited to the weddings etc... In other words, I trust them, they are good folks. Regardless, we would also want to get a kvm for the machines. Joshua D. Drake > > //Magnus > > ---------------------------(end of broadcast)--------------------------- > TIP 9: In versions below 8.0, the planner will ignore your desire to > choose an index scan if your joining column's datatypes do not > match >
On Thu, Aug 02, 2007 at 04:58:34PM -0700, Joshua D. Drake wrote: > Oh and just so everyone understands the relationship with the data > center, I started it... years before CMD existed. I drink beer and > golf with the owners, I am invited to the weddings etc... That's quite adventurous. What does golf taste like? ;) > In other words, I trust them, they are good folks. > > Regardless, we would also want to get a kvm for the machines. KVM/IP is awesome. I'm told that some of the really good ones let you use local media (CD, DVD, etc.) to boot the remote machine. Cheers, D -- David Fetter <david@fetter.org> http://fetter.org/ phone: +1 415 235 3778 AIM: dfetter666 Skype: davidfetter Remember to vote! Consider donating to PostgreSQL: http://www.postgresql.org/about/donate
David Fetter wrote:
> On Thu, Aug 02, 2007 at 04:58:34PM -0700, Joshua D. Drake wrote:
>> Oh and just so everyone understands the relationship with the data
>> center, I started it... years before CMD existed. I drink beer and
>> golf with the owners, I am invited to the weddings etc...
>
> That's quite adventurous. What does golf taste like? ;)
Bloody Mary ;)
>
>> In other words, I trust them, they are good folks.
>>
>> Regardless, we would also want to get a kvm for the machines.
>
> KVM/IP is awesome. I'm told that some of the really good ones let you
> use local media (CD, DVD, etc.) to boot the remote machine.
Yes they do :)
Joshua D. Drake
>
> Cheers,
> D
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
David Fetter wrote: > On Thu, Aug 02, 2007 at 04:58:34PM -0700, Joshua D. Drake wrote: >> Oh and just so everyone understands the relationship with the data >> center, I started it... years before CMD existed. I drink beer and >> golf with the owners, I am invited to the weddings etc... > > That's quite adventurous. What does golf taste like? ;) Oh thats a gem - I nearly fell off my chair! /D
Joshua D. Drake wrote: > We have two people that are CMD people that are local to the machines. > However the data center has people in house (not just on call but 20 > feet away) and any thing that needs to be done 8-6PST could be done via > phone.> > It is actually how CMD does *everything* now with the exception of > provisioning which I reserve for myself to make sure everything stays sane. OK, then I have no objection to you guys hosting these machines. Regards, Dave.
On Fri, Aug 03, 2007 at 08:42:20AM +0100, Dave Page wrote: > Joshua D. Drake wrote: > >We have two people that are CMD people that are local to the machines. > >However the data center has people in house (not just on call but 20 > >feet away) and any thing that needs to be done 8-6PST could be done via > >phone. > > > >It is actually how CMD does *everything* now with the exception of > >provisioning which I reserve for myself to make sure everything stays sane. > > OK, then I have no objection to you guys hosting these machines. Agreed from here. //Magnus
Folks, So, we're all good that barring objections from equipment/funding donors, we'll rack our equipment at CommandPrompt and list them as a contributor to the lab? Josh, can you figure out what I need to raise in power expenses per year? -- Josh Berkus PostgreSQL @ Sun San Francisco
Josh Berkus wrote:
> Folks,
>
> So, we're all good that barring objections from equipment/funding donors,
> we'll rack our equipment at CommandPrompt and list them as a contributor to
> the lab?
>
> Josh, can you figure out what I need to raise in power expenses per year?
It won't be much and actually will depend on the "type" of machines
because we may have to pull in power. E.g; right now we are wired for
110, but some bigger machines will require 220.
It isn't a big deal either way as we have the power available (I
actually spoke with them last night when I brought up the new pgfoundry).
As a note, the power is all conditioned and backed by natural gas generator.
Joshua D. Drake
>
--
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997 http://www.commandprompt.com/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
Josh, > It won't be much and actually will depend on the "type" of machines > because we may have to pull in power. E.g; right now we are wired for > 110, but some bigger machines will require 220. Yeah, I think some of the Sun machines I've asked for are 220. What's the sort of cost we're looking at for pulling in power? We don't want to do this with CMD if it's going to be *more* expensive than OSL. -- Josh Berkus PostgreSQL @ Sun San Francisco
Folks, We never made a final decision on the CMD vs. OSL decision on where to house these. I'll need to ship some boxes this week, so it would be great if we could settle this. Josh, what's our power cost likely to be with you? -- --Josh Josh Berkus PostgreSQL @ Sun San Francisco
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Josh Berkus wrote:
> Folks,
>
> We never made a final decision on the CMD vs. OSL decision on where to
> house these. I'll need to ship some boxes this week, so it would be great
> if we could settle this.
>
> Josh, what's our power cost likely to be with you?
I don't have an exact number and wasn't going to worry about it until it
became an issue. I believe the comparative was about 2k a year which we
can beat as I currently have quite the surplus on power.
Sincerely,
Joshua D. Drake
- --
=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 24x7/Emergency: +1.800.492.2240
PostgreSQL solutions since 1997 http://www.commandprompt.com/ UNIQUE NOT NULL
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFGygb8ATb/zqfZUUQRArgpAJsHMKSyL16B2zcHVxUaFk/5KI2FgACfV9DU
NsUbSG2dEYSP97XeE/fqf0M=
=3oMk
-----END PGP SIGNATURE-----
Joshua D. Drake wrote: > Josh Berkus wrote: >> Folks, > >> We never made a final decision on the CMD vs. OSL decision on where to >> house these. I'll need to ship some boxes this week, so it would be great >> if we could settle this. > >> Josh, what's our power cost likely to be with you? > > I don't have an exact number and wasn't going to worry about it until it > became an issue. I believe the comparative was about 2k a year which we > can beat as I currently have quite the surplus on power. 2K per year!!! What the heck is Josh shipping - a Cray XMP? /D
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 - --On Monday, August 20, 2007 22:37:08 +0100 Dave Page <dpage@postgresql.org> wrote: > Joshua D. Drake wrote: >> Josh Berkus wrote: >>> Folks, >> >>> We never made a final decision on the CMD vs. OSL decision on where to >>> house these. I'll need to ship some boxes this week, so it would be great >>> if we could settle this. >> >>> Josh, what's our power cost likely to be with you? >> >> I don't have an exact number and wasn't going to worry about it until it >> became an issue. I believe the comparative was about 2k a year which we >> can beat as I currently have quite the surplus on power. > > 2K per year!!! What the heck is Josh shipping - a Cray XMP? - From how Joshua worded it about, my take was that $2k was what OSL quoted JoshB ... but, with OSL, was that 'just power', or was that 'power and bandwidth'? - ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . scrappy@hub.org MSN . scrappy@hub.org Yahoo . yscrappy Skype: hub.org ICQ . 7615664 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4 (FreeBSD) iD8DBQFGygsC4QvfyHIvDvMRAjsAAKDn5J2OHkFas4Q4cZlyoS2FiYeA6ACfcE/4 1LupmQKnKfszjIj2BSYkEd8= =/geT -----END PGP SIGNATURE-----
Marc, Josh, > From how Joshua worded it about, my take was that $2k was what OSL > quoted JoshB > ... but, with OSL, was that 'just power', or was that 'power and > bandwidth'? That was "we'd have to donate a couple $thousand to cover power & help getting our rack set up." -- --Josh Josh Berkus PostgreSQL @ Sun San Francisco