Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 4 Dec 2017 00:31:14 +0000
From:      Rick Macklem <rmacklem@uoguelph.ca>
To:        Colin Percival <cperciva@tarsnap.com>, Rafal Lukawiecki <raf@rafal.net>, "freebsd-cloud@freebsd.org" <freebsd-cloud@freebsd.org>
Subject:   Re: Suppress kernel messages "NFSv4 fileid > 32bits"
Message-ID:  <YTOPR0101MB2172DD382234F70C201C7654DD3C0@YTOPR0101MB2172.CANPRD01.PROD.OUTLOOK.COM>
In-Reply-To: <010001601e629361-9732ac13-ac59-462a-b30f-92702c3600a4-000000@email.amazonses.com>
References:  <386518E3-1D01-4E1F-BB77-E9C530E05381@rafal.net> <0100015ff59bd564-84bd069a-00b5-4a05-a8aa-dbe2aa19693c-000000@email.amazonses.com> <D64EB3EA-37B0-4FC3-9D0C-31642030B654@rafal.net>, <010001601e629361-9732ac13-ac59-462a-b30f-92702c3600a4-000000@email.amazonses.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Colin Percival wrote:
>On 12/03/17 09:18, Rafal Lukawiecki wrote:
>> Is there a way to completely suppress the "NFSv4 fileid > 32bits=94 mess=
ages
>> from flooding my logs and the subsequent emailed reports? This message i=
s
>> unavoidable when running on AWS until FreeBSD 12 comes out, based on the
>> previous discussions here.
>
>Well, unavoidable when using EFS.  Most people in EC2 aren't using EFS.
>
>> I can see that Colin has worked on reducing their number
>> (https://reviews.freebsd.org/D8523) but I would prefer to disable this
>> message completely.
>
>This is more of an NFS question than an EC2 question, but sure, suppressin=
g
>that message completely is entirely doable.
>
>CCing in Rick, as the maintainer of the NFS code -- Rick, any opinion on
>having a vfs.nfs.suppress_32bits_warning sysctl, defaulting to 0, which
>suppresses those two warnings when set to nonzero?
My only suggestion is that the message should be generated at least once,
since it does indicate that your mount is "broken". When the fileid (i-node=
#)
is truncated to 32bits, it is possible for two different fileids to be mapp=
ed to
the same number. The most common error I am aware of is the fts(3) library
functions generating "cycle in directory structure" errors.

It would have to be a direct commit to stable/11, but if Colin wants to do =
it,
I don't have a problem with it, rick




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?YTOPR0101MB2172DD382234F70C201C7654DD3C0>