lkml.org 
[lkml]   [1998]   [Nov]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: NFS read performance ugliness
Hi,

Zlatko Calusic:
>
> I'm attaching two tcpdump logs (one from Solaris and one from
> Linux). On both logs, machine names are shortened to local (always
> Linux - client) and remote (always Solaris - server). Logs are taken
> while I was doing cp /dev/zero, on Linux side (copying to NFS mounted
> directory, of course).
>
> 1/2) tcpdump log on Solaris (server) side:
> [...]
> 2/2) tcpdump log on Linux (client) side
>
> 11:37:23.484132 local > remote: (frag 1251:1308@2960)
> 11:37:23.484227 local > remote: (frag 1251:1480@1480+)
> 11:37:23.484343 local.4047701504 > remote.nfs: 1472 write [|nfs] (frag 1251:1480@0+)
> 11:37:23.542237 remote.nfs > local.4047701504: reply ok 96 write [|nfs] (DF)

That looks like the Solaris box takes its time writing the data to disk
(i.e., no networking problems).

I don't have any idea why it would do that. 50 msec for a simple 4096-byte
write?? Give me a break...

--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
You have several detached dead trees.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:45    [W:0.056 / U:0.700 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site