Yeah those big disks arrays are real sweet.
One day last week I was in a data center in Arizona when the big LSI/Storagetek
array in the cage next to mine had a hard drive failure. So the alarm shrieked
at like 13225535 decibles continuously for hours. BEEEP BP BP BP.
Of course sinc
Quoting Lane Van Ingen <[EMAIL PROTECTED]>:
> ... to do the following:
> (1) Make a table memory-resident only ?
Put it on a RAM filesystem. On Linux, shmfs. On *BSD, mfs. Solaris, tmpfs.
> (2) Set up user variables in memory that are persistent across all
> sessions, for
> as long
Bear in mind you will lose data if the raid controller itself fails (or the
cache memory module). Many solutions have mirrored cache for this reason. But
that's more $$, depending on the risks you want to take.
Quoting Arjen van der Meijden <[EMAIL PROTECTED]>:
> On 14-9-2005 22:03, Alvaro Herr
This might be optimal behavior from the hardware. Random reads are hard to
optimize for--except if you have enough physical memory to hold the entire
dataset. Cached reads (either in array controller or OS buffer cache) should
return nearly immediately. But random reads probably aren't cached.
Quoting Barry Lind <[EMAIL PROTECTED]>:
>
>
> What I see when running the V3 protocol under 'top' is that the postgres
> processes are routinely using 15% or more of the CPU each, when running
> the V2 protocol they use more like 0.3%.
>
>
>
> Does anyone have any suggestions on an approac
Quoting Anjan Dave <[EMAIL PROTECTED]>:
> Hi,
>
>
>
> One simple question. For 125 or more checkpoint segments
> (checkpoint_timeout is 600 seconds, shared_buffers are at 21760 or
> 170MB) on a very busy database, what is more suitable, a separate 6 disk
> RAID5 volume, or a RAID10 volume? Dat
Quoting "Jeffrey W. Baker" <[EMAIL PROTECTED]>:
>
> Here's the result, in transactions per second.
>
> ext3 jfs xfs
> --
---
> 10 Clients 55 81 68
> 100 Clients 61 100 64
>
Was fsync true? And have you tri
If you truly do not care about data protection -- either from drive loss or from
sudden power failure, or anything else -- and just want to get the fastest
possible performance, then do RAID 0 (striping). It may be faster to do that
with software RAID on the host than with a special RAID controlle
Thanks, Andrew. I expect to choose between HBAs with no RAID functionality or
with the option to completely bypass RAID functionality--meaning that I'll
hopefully avoid the situation that you've described. I'm mostly curious as to
whether the driver problems described for U320 Adaptec RAID contro
I have a similar question about what to choose (either LSI or Adaptec U320), but
plan to use them just for JBOD drivers. I expect to be using either net or
freebsd. The system CPU will be Opteron. My impression is that both the ahd
and mpt drivers (for U320 Adaptec and LSI, respectively) are qui
Quoting Randolf Richardson <[EMAIL PROTECTED]>:
> I'm looking for recent performance statistics on PostgreSQL vs. Oracle
>
> vs. Microsoft SQL Server. Recently someone has been trying to convince my
I don't know anything about your customer's requirements other than that they
have a DB c
I fully agree with Gaetano about testing sync methods. From testing I've done
on two different Solaris 8 boxes, the O_DSYNC option on Solaris 8 beats fsync
and fdatasync easily. Test it yourself though. There's probably some
opportuntiy there for better performance for you.
> > BTW this is Post
Quoting Josh Berkus <[EMAIL PROTECTED]>:
> The first and most important step for RAID performance with PostgreSQL is to
>
> get a card with onboard battery back-up and enable the write cache for the
> card. You do not want to enable the write cache *without* battery back-up
>
I'm curious abo
Quoting Bruce Momjian <[EMAIL PROTECTED]>:
>
> Have you seen /src/tools/fsync?
>
I have now. Thanks.
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
Hi, I'd like to help with the topic in the Subject: line. It seems to be a
TODO item. I've reviewed some threads discussing the matter, so I hope I've
acquired enough history concerning it. I've taken an initial swipe at
figuring out how to optimize sync'ing methods. It's based larg
15 matches
Mail list logo
