Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 28 Dec 1999 01:10:01 -0800 (PST)
From:      Kevin Day <toasty@dragondata.com>
To:        freebsd-bugs@FreeBSD.org
Subject:   Re: kern/7619
Message-ID:  <199912280910.BAA12457@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/7619; it has been noted by GNATS.

From: Kevin Day <toasty@dragondata.com>
To: FreeBSD-gnats-submit@FreeBSD.org
Cc:  
Subject: Re: kern/7619
Date: Tue, 28 Dec 1999 03:03:16 -0600 (CST)

 Just for the audit trail:
 
 
 This turned out to be the NFS client dynamic retransmitter getting overly
 aggressive when it came to very very fast links (like 100MB Ethernet). The
 normal rtt is very short, except when there are a few collisions, which
 would dramatically increase the rtt.
 
 Mounting with '-d' will correct this problem, and I highly recommend anyone
 who's seeing the 'server not responding' messages in their syslog to use it.
 The NFS client will tend to get rather confused when it sees the server go
 away, and will sometimes get rather out of sync.
 
 After looking at the rather scary code used for the dynamic retransmit, I've
 decided against trying to fix it myself. :)
 
 #define NFS_RTO(n, t) \
         ((t) == 0 ? (n)->nm_timeo : \
          ((t) < 3 ? \
           (((((n)->nm_srtt[t-1] + 3) >> 2) + (n)->nm_sdrtt[t-1] + 1) >> 1) : \
           ((((n)->nm_srtt[t-1] + 7) >> 3) + (n)->nm_sdrtt[t-1] + 1)))
 
 I do not wish to break this. :)
 
 (Search the archives around April/May in -current for a discussion about
 this... I think someone else looked at the retransmit stuff and came to the
 same conclusion)
 
 Kevin
 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199912280910.BAA12457>