From: Doug Nazar <nazard@nazar.ca>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: linux-nfs@vger.kernel.org
Subject: Re: [PATCH 11/11] svcgssd: Wait for nullrpc channel if not available
Date: Sat, 18 Jul 2020 14:07:37 -0400 [thread overview]
Message-ID: <3b9a70ce-f1b4-da68-cfc1-8e93af197461@nazar.ca> (raw)
In-Reply-To: <20200718155501.GB27817@fieldses.org>
On 2020-07-18 11:55, J. Bruce Fields wrote:
> So, is there a race here that could result in a hang, and has anyone
> seen it in practice?
>
> Just curious. Thanks for doing this.--b.
Not a hang, with the existing code, svcgssd will just exit. I'd have to
go and restart it after boot. I'm assuming a systemd setup would just
restart it automatically.
On my original systems I'd configured it to force load the modules, but
I'd forgotten about that (it was about 10 years ago) when I built this
new box last month.
As mentioned in the other email, sunrpc is loaded from an initramfs that
doesn't have any of the gss modules. When nfsd is started (after
svcgssd) it'll then load the gss modules but by then it's too late.
It seems to be a standard Gentoo setup. I just checked the default
kernel config for genkernel (their kernel & initramfs builder) and it
has all the rpc & nfs as modules, with the initramfs not supporting
Kerberos.
Looks like Debian modprobes rpcsec_gss_krb5 before starting rpc.svcgssd.
So with this waiting for the file to appear, and I'm planning on adding
a conditional module load to the Gentoo rpc.svcgssd init script, things
should be as bulletproof as I can make it.
Doug
next prev parent reply other threads:[~2020-07-18 18:07 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-18 9:24 [PATCH 00/11] nfs-utils: Misc cleanups & fixes Doug Nazar
2020-07-18 9:24 ` [PATCH 01/11] Add error handling to libevent allocations Doug Nazar
2020-07-18 9:24 ` [PATCH 02/11] gssd: Fix cccache buffer size Doug Nazar
2020-07-20 14:43 ` Steve Dickson
2020-07-20 15:41 ` Doug Nazar
2020-07-18 9:24 ` [PATCH 03/11] gssd: Fix handling of failed allocations Doug Nazar
2020-07-18 9:24 ` [PATCH 04/11] gssd: srchost should never be * Doug Nazar
2020-07-18 9:24 ` [PATCH 05/11] xlog: Reorganize xlog_backend() to work around -Wmaybe-uninitialized Doug Nazar
2020-07-18 9:24 ` [PATCH 06/11] nfsdcld: Add graceful exit handling and resource cleanup Doug Nazar
2020-07-18 9:24 ` [PATCH 07/11] nfsdcld: Don't copy more data than exists in column Doug Nazar
2020-07-18 9:24 ` [PATCH 08/11] svcgssd: Convert to using libevent Doug Nazar
2020-07-18 9:24 ` [PATCH 09/11] nfsidmap: Add support to cleanup resources on exit Doug Nazar
2020-07-20 15:49 ` Steve Dickson
2020-07-20 15:58 ` Doug Nazar
2020-07-20 17:31 ` Steve Dickson
2020-07-18 9:24 ` [PATCH 10/11] svcgssd: Cleanup global " Doug Nazar
2020-07-18 9:24 ` [PATCH 11/11] svcgssd: Wait for nullrpc channel if not available Doug Nazar
2020-07-18 15:55 ` J. Bruce Fields
2020-07-18 18:07 ` Doug Nazar [this message]
2020-07-27 14:42 ` [PATCH 00/11] nfs-utils: Misc cleanups & fixes Steve Dickson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3b9a70ce-f1b4-da68-cfc1-8e93af197461@nazar.ca \
--to=nazard@nazar.ca \
--cc=bfields@fieldses.org \
--cc=linux-nfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).