Thread: error: C preprocessor "/lib/cpp" fails sanity check

error: C preprocessor "/lib/cpp" fails sanity check

From
AI Rumman
Date:
I am getting the following error during Postgresql 8.1 installation:
 
error: C preprocessor "/lib/cpp" fails sanity check
 
Please any suggestion how to solve it.

Re: error: C preprocessor "/lib/cpp" fails sanity check

From
John R Pierce
Date:
AI Rumman wrote:
> I am getting the following error during Postgresql 8.1 installation:
>
> error: C preprocessor "/lib/cpp" fails sanity check
>
> Please any suggestion how to solve it.

what platform (os, version, etc) are you building postgres on?

what C++ compiler is that?   thats an unusual path for a C++ compiler,
/usr/bin/cpp is more typical.



SAS Raid10 vs SATA II Raid10 - many small reads and writes

From
Phillip Berry
Date:
Hi Everyone,

We're in the market for a new DB server to replace our current one (yes it's one of *those*
questions) ;).

It'll have quad core Xeons, 36GB RAM and some sort of Raid 10 configuration.

Our provider is pushing us towards 6 x SATA II disks in a Raid 10 configuration or 4 x SAS disks in
Raid 10 (budget constraints).

The application that queries this DB opens up about 100 connections and performs millions of
inherently small reads and writes to process the data over the course of a few weeks.

My question is though the maximum throughput may be similar between 6 SATA II and 4 SAS drives does
anybody know if the SAS drives will *significantly* outperform the SATA drives in the face of many
millions of small concurrent reads and writes?

My Google-fu is weak today and I can't seem to come up with a definitive answer one way or the other,
does anyone have any experience that they can offer?

Kind Regards
Phil Berry


Re: error: C preprocessor "/lib/cpp" fails sanity check

From
Peter Eisentraut
Date:
On ons, 2010-03-10 at 12:04 +0600, AI Rumman wrote:
> I am getting the following error during Postgresql 8.1 installation:
>
> error: C preprocessor "/lib/cpp" fails sanity check
>
> Please any suggestion how to solve it.

Check the config.log file.  I would guess it's somewhat likely that you
don't have a full C compiler suite installed at all, or the installation
is botched.


Re: SAS Raid10 vs SATA II Raid10 - many small reads and writes

From
Scott Marlowe
Date:
On Tue, Mar 9, 2010 at 11:49 PM, Phillip Berry
<pberry@stellaconcepts.com> wrote:
> Hi Everyone,
>
> We're in the market for a new DB server to replace our current one (yes it's one of *those*
> questions) ;).
>
> It'll have quad core Xeons, 36GB RAM and some sort of Raid 10 configuration.
>
> Our provider is pushing us towards 6 x SATA II disks in a Raid 10 configuration or 4 x SAS disks in
> Raid 10 (budget constraints).

Are those your only two options?  No 6 SAS drives?  Are you looking at
7200rpm or 10krpm SATA?  15krpm or 10krpm SAS?  What RAID controller?
Battery backed Cache?  Software RAID?

Re: SAS Raid10 vs SATA II Raid10 - many small reads and writes

From
Phillip Berry
Date:
On Wednesday 10 March 2010 18:32:41 Scott Marlowe wrote:
> On Tue, Mar 9, 2010 at 11:49 PM, Phillip Berry
>
> <pberry@stellaconcepts.com> wrote:
> > Hi Everyone,
> >
> > We're in the market for a new DB server to replace our current one (yes
> > it's one of *those* questions) ;).
> >
> > It'll have quad core Xeons, 36GB RAM and some sort of Raid 10
> > configuration.
> >
> > Our provider is pushing us towards 6 x SATA II disks in a Raid 10
> > configuration or 4 x SAS disks in Raid 10 (budget constraints).
>
> Are those your only two options?  No 6 SAS drives?  Are you looking at
> 7200rpm or 10krpm SATA?  15krpm or 10krpm SAS?  What RAID controller?
> Battery backed Cache?  Software RAID?
>

They're charging  a lot more for SAS than for SATA so it's a budget constraint. And we can only fit
six drives in the machine so that limits the number of drives for the SATA option.

Hardware raid controller for both options, but I'm not sure what brand (yet).

15,000rpm SAS
10,000rpm SATA


As with everything it's a trade off: 4 SAS drives or 6 SATA II drives in raid 10. I'm trying to find
out if (for many many small reads and writes) one is more desirable than the other.

Regards
Phillip Berry


Re: SAS Raid10 vs SATA II Raid10 - many small reads and writes

From
Magnus Hagander
Date:
On Thursday, March 11, 2010, Phillip Berry <pberry@stellaconcepts.com> wrote:
> On Wednesday 10 March 2010 18:32:41 Scott Marlowe wrote:
>> On Tue, Mar 9, 2010 at 11:49 PM, Phillip Berry
>>
>> <pberry@stellaconcepts.com> wrote:
>> > Hi Everyone,
>> >
>> > We're in the market for a new DB server to replace our current one (yes
>> > it's one of *those* questions) ;).
>> >
>> > It'll have quad core Xeons, 36GB RAM and some sort of Raid 10
>> > configuration.
>> >
>> > Our provider is pushing us towards 6 x SATA II disks in a Raid 10
>> > configuration or 4 x SAS disks in Raid 10 (budget constraints).
>>
>> Are those your only two options?  No 6 SAS drives?  Are you looking at
>> 7200rpm or 10krpm SATA?  15krpm or 10krpm SAS?  What RAID controller?
>> Battery backed Cache?  Software RAID?
>>
>
> They're charging  a lot more for SAS than for SATA so it's a budget constraint. And we can only fit
> six drives in the machine so that limits the number of drives for the SATA option.
>
> Hardware raid controller for both options, but I'm not sure what brand (yet).
>
> 15,000rpm SAS
> 10,000rpm SATA
>
>
> As with everything it's a trade off: 4 SAS drives or 6 SATA II drives in raid 10. I'm trying to find
> out if (for many many small reads and writes) one is more desirable than the other.


How much do you have to cut your CPU and/or memory (CPU first!) to get
to 6 SAS? It may be well worth considering....


--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

Re: SAS Raid10 vs SATA II Raid10 - many small reads and writes

From
Greg Smith
Date:
Phillip Berry wrote:
> Hardware raid controller for both options, but I'm not sure what brand
> (yet).

The reason you won't ever find a good general answer to this question is
that it's so close that you need to know the exact controller cards and
the disks used in each situation to have any hope of guessing which will
work out better.  Sometimes you get lucky and one case has a know poor
performer, so you just get the other.  Next tiebreak is management
utilites, which count for a lot more than a small performance
difference.  If both are good on both of those, then maybe it's worth
the trouble to model each possibility against your workload, with
accurate numbers to substitute into any such guess.

--
Greg Smith  2ndQuadrant US  Baltimore, MD
PostgreSQL Training, Services and Support
greg@2ndQuadrant.com   www.2ndQuadrant.us