Date: Tue, 3 Jul 2001 17:51:50 -0400 From: "Deepak Jain" <deepak@ai.net> To: <questions@FreeBSD.ORG> Subject: RE: TX buffer in 4.3 Message-ID: <GPEOJKGHAMKFIOMAGMDIEEFIDLAA.deepak@ai.net> In-Reply-To: <20010703161057.O47870@elvis.mu.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Bill's explanation here is very thorough, but unless you know what he means
by the last comment, it won't necessarily help you.
What he is suggesting is adding a line to your kernel config,
options NMBCLUSTERS=2048
(for example)
and loading the new kernel. This should solve your transmit issues on the
server in question.
The FreeBSD handbook probably tells you how to get the kernel sources if you
don't have them and build the new kernel.
If you already know how to do this, then I am just mentioning it for
pedantic concerns.
Deepak Jain
AiNET
-----Original Message-----
From: owner-freebsd-hackers@FreeBSD.ORG
[mailto:owner-freebsd-hackers@FreeBSD.ORG]On Behalf Of Bill Fumerola
Sent: Tuesday, July 03, 2001 5:11 PM
To: Randy -Harborside Internet
Cc: freebsd-hackers@FreeBSD.ORG
Subject: Re: TX buffer in 4.3
[ not -hackers material, moved to questions ]
On Tue, Jul 03, 2001 at 01:53:43PM -0700, Randy -Harborside Internet wrote:
> We are having a problem with our mail server. It recently got
> upgraded to 4.3 from 4.2, and now it is having problems with the TX
> buffer somehow on the network card. Every once in awhile it will shut
> off all network traffic and give these errors:
> no memory for tx listrl0
> Then in a few minutes (presumably when the buffer is flushed somehow)
> the network device resumes normal operation.
s#flushed#drained#
> We have tried 3 different NICs, and all have had the same problem.
> The three models were:
> 1. Realtek RTL8139A 10/100TX
rl(4):
rl%d: no memory for tx list The driver failed to allocate an mbuf for
the transmitter ring when allocating a pad buffer or collapsing
an mbuf
chain into a cluster.
> 2. Intel chipset:82558B
fxp(4):
fxp%d: Failed to malloc memory There are not enough mbuf's available
for
allocation.
> 3. 3Com somethingerather.
I'll assume you mean the etherlink(3c905) chips...
xl(4):
xl%d: no memory for tx list The driver failed to allocate an mbuf for
the transmitter ring when allocating a pad buffer or collapsing
an mbuf
chain into a cluster.
> We are running with 512MB of RAM, and it usually has about 200 or
> more megs free at the time of this occurance.
Which doesn't matter if you don't allocate enough memory to the mbuf
subsystem.
See below.
> Is this a problem with the network drivers in 4.3? Or something else
> that can be corrected? (Manual way to flush the network card
> buffers??)
Flush the network card buffers? That really wouldn't help anything.
> Here is the output of ulimit -a, just in case that helps.
It doesn't.
If you look at the output of 'netstat -m' you'll see you've run mbufs
(well, mbuf clusters) and that is whats causing this problem. The
"requests for memory" lines will show you how many times this has bitten
you in the proverbial ass.
options NMBCLUSTERS is your friend.
--
Bill Fumerola - security yahoo / Yahoo! inc.
- fumerola@yahoo-inc.com / billf@FreeBSD.org
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?GPEOJKGHAMKFIOMAGMDIEEFIDLAA.deepak>