Hi guys. I'm sending you guys a link for a somewhat complex I/O
benchmark I wrote for doing all kinds of noncontiguous I/O through
MPI-IO. It's called HPIO and does all kinds of weird tests. Anyway, it
is really overkill for this small I/O bug, but the nice thing is that I
have a verify mode whi
nevermind - i'm not reading carefully. -- rob
Robert Latham wrote:
On Wed, Feb 22, 2006 at 03:22:35PM -0600, Avery Ching wrote:
Sure. I'm actually have a single client just doing a small contiguous
write of 50 bytes. But I think it occurs for pretty much any small I/O
operation I do. It seem
small i/o is only used in contig/contig cases, so the test you mention
isn't going to trigger what avery is describing.
rob
Robert Latham wrote:
On Wed, Feb 22, 2006 at 03:22:35PM -0600, Avery Ching wrote:
Sure. I'm actually have a single client just doing a small contiguous
write of 50 byte
Hi David,
I tried to reproduce your results with the 2.6 kernel, and wasn't
able to. Are you using 2.4? Also, I didn't actually pull the plug
on one of the nodes, I just killed the server, but that should be
close enough to your test case unless you're routing stuff through
that node ;
On Wed, Feb 22, 2006 at 03:22:35PM -0600, Avery Ching wrote:
> Sure. I'm actually have a single client just doing a small contiguous
> write of 50 bytes. But I think it occurs for pretty much any small I/O
> operation I do. It seems to happen with any amount of PVFS2 servers.
> Both the memory r
Sure. I'm actually have a single client just doing a small contiguous
write of 50 bytes. But I think it occurs for pretty much any small I/O
operation I do. It seems to happen with any amount of PVFS2 servers.
Both the memory request and file request structures are contig I think.
The total_comp
Hi Avery,
Can you let me know what version of pvfs2 you're using, and also the
patterns of the reads you're doing (the memory request and file
request structures passed to sys_read)?
Thanks,
-sam
On Feb 22, 2006, at 2:05 PM, Avery Ching wrote:
Hi guys,
I've been trying to debug
Hi guys,
I've been trying to debug a nasty noncontiguous I/O problem in PVFS2
and noticed a problem with the small I/O case. It appears that the
resp_io.total_completed = 0 in the write case even though some data
seems to be written to the file. I was thinking it might be because the
smal
> -Original Message-
> From: Sam Lang [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, February 22, 2006 12:47 PM
> To: [EMAIL PROTECTED]
> Cc: pvfs2-developers@beowulf-underground.org
> Subject: Re: [Pvfs2-developers] Problem with multiple pvfs2
> file systems mounted on a single client
>
>
Hi David,
Just to clarify your setup a bit, you are running two separate sets
of metadata and io servers for the two different mountpoints, so that
for /mnt/pvfs2 you have one set of nodes running the servers, and
for /mnt/pvfs2-tmp you have a completely separate set of server
nodes? An
It appears the error described below will span across other mounted file
systems on a client when encountered, until the client software is reloaded.
I've got a client with 2 pvfs2 file systems mounted:
/mnt/pvfs2
/mnt/pvfs2-tmp
Both PVFS2 file system configurations contained th
[EMAIL PROTECTED] wrote on Wed, 22 Feb 2006 10:38 -0600:
> Either name would be fine; PINT_ is a little better I suppose.
Thanks, I'll rename them as such and move the macros up above BMI
once we find a couple more uses for them (openib at least).
-- Pete
_
Either name would be fine; PINT_ is a little better I suppose.
Rob
Pete Wyckoff wrote:
[EMAIL PROTECTED] wrote on Wed, 22 Feb 2006 10:10 -0500:
It appears that compile errors I'm running into with PVFS2 and
Silverstorm/Mellanox infiniband are definitions in the ib code that
duplicate defini
Assert is also redefined in that same ib file. So we may think about
changing that one also if the def from assert.h won't cut it, or at
least using a #undefine there.
I also have to tweak the iba headers, but that seems to be limited to
just adding a const to make so messages go away.
Brad
[EMAIL PROTECTED] wrote on Wed, 22 Feb 2006 10:10 -0500:
> It appears that compile errors I'm running into with PVFS2 and
> Silverstorm/Mellanox infiniband are definitions in the ib code that
> duplicate definitions in kernel headers. Specifically:
>
> #define likely(x) __builtin_expect
[EMAIL PROTECTED] wrote on Tue, 21 Feb 2006 16:56 -0600:
> According to the profiling, we're looking at roughly 50% of total
> cpu-time being spent inside check_cq() in a given run.
> ( src/io/bmi/bmi_ib/ib.c:check_cq() )
> --I'll further test this on some longer tests tomorrow to get more
> acc
Pete,
It appears that compile errors I'm running into with PVFS2 and
Silverstorm/Mellanox infiniband are definitions in the ib code that
duplicate definitions in kernel headers. Specifically:
#define likely(x) __builtin_expect(!!(x), 1)
#define unlikely(x) __builtin_expect(!!(x),
Hello.
im working with pvfs2 and im reading to the doc (FAQ y pvfs2 guide) and i
dont found the site in src code where is round-robin policy for IO
operations because i would change this policy.
thanks
___
Pvfs2-developers mailing list
Pvfs2-developers
18 matches
Mail list logo