linux-cifs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Murphy Zhou <jencce.kernel@gmail.com>
To: Steve French <smfrench@gmail.com>
Cc: "Murphy Zhou" <jencce.kernel@gmail.com>,
	"ronnie sahlberg" <ronniesahlberg@gmail.com>,
	"Aurélien Aptel" <aaptel@suse.com>,
	CIFS <linux-cifs@vger.kernel.org>
Subject: Re: Are the xfstests exclusion files on wiki.samba.org up to date?
Date: Sat, 14 Sep 2019 08:44:51 +0800	[thread overview]
Message-ID: <20190914004451.k2fhirxo54hgh5ln@XZHOUW.usersys.redhat.com> (raw)
In-Reply-To: <CAH2r5mv5RqAfAMb2oWa+10bqUa+6kUMO3HKPG1s2z0PRqOeKEg@mail.gmail.com>

On Fri, Sep 13, 2019 at 05:03:57PM -0700, Steve French wrote:
> Actually the xfstests we pass compared favorably with other network and
> cluster fs but just different :)

Agreed :)

M
> 
> On Fri, Sep 13, 2019, 17:02 Murphy Zhou <jencce.kernel@gmail.com> wrote:
> 
> > On Thu, Sep 12, 2019 at 07:06:52PM +1000, ronnie sahlberg wrote:
> > > On Thu, Sep 12, 2019 at 6:50 PM Murphy Zhou <jencce.kernel@gmail.com>
> > wrote:
> > > >
> > > > On Mon, Sep 09, 2019 at 06:23:10PM -0500, Steve French wrote:
> > > > > We have done a lot of work (with Ronnie, Aurelien, Pavel and others)
> > > > > on xfstest automation.
> > > >
> > > > Kudos!
> > > >
> > > > >
> > > > > As you might guess it is frustrating for two reasons:
> > > > > 1) updated xfstests can be flaky (as the tests themselves are
> > updated,
> > > > > they add subtle required features, or regress from time to time)
> > > > > 2) test automation can run into resource constraints (VMs trying to
> > > > > run tests with less memory than might be optimal - especially those
> > > > > run against Samba with "VMs inside VMs").
> > > > >
> > > > > But the good news is that we have VERY good data on which tests pass
> > > > > to various servers (just as noted, need to update the external pages)
> > > > > and we should have even more data as we go through two weeks of SMB3
> > > > > testing events in late September.
> > > > >
> > > > > In general we want to test against all typical servers and have test
> > > > > targets setup for
> > > > >  - Samba (reasonably current stable) on ext4/xfs and Samba with btrfs
> > > > > (which has various optional extensions enabled in the server)
> > > > >  - Samba with POSIX Extensions
> > > > >  - Windows (and against both NTFS and REFS server file system)
> > > > > - Azure (Cloud0
> > > >
> > > > Thanks for sharing this!
> > > >
> > > > Besides server type, server configuration options and client mount
> > > > options also make this matrix even bigger..
> > > >
> > > > How that is handled in the buildbot ?
> > >
> > > We have several different targets, which maps to the local.config for
> > > xfstests so for every single test we run
> > > we also specify the different targets to run that for.
> > >
> > > Here is an example:
> > > all_tests = [
> > >         [ "cifs/100", "smb3azureseal"],
> > >         [ "cifs/101", "smb3multiuser"],
> > >         [ "generic/001", "smb3sign", "smb3"],
> > >         [ "generic/002", "smb3", "smb3sign", "smb21", "smb3samba"],
> > >         [ "generic/005", "smb3", "smb21", "smb3samba", "smb3sambabtrfs"],
> > >         [ "generic/006", "smb3"],
> > >         [ "generic/007", "smb3"],
> > >         [ "generic/010", "smb3"],
> > >         [ "generic/011", "smb3"],
> > >         [ "generic/013", "smb3samba"],
> > >         [ "generic/014", "smb3"],
> > >         [ "generic/020", "smb3samba"],
> > >         [ "generic/023", "smb3samba"],
> > >         [ "generic/024", "smb3", "smb3samba"],
> > >         [ "generic/028", "smb3", "smb3samba"],
> > > ...
> > >
> > >
> > > The first argument is which xfstest to run and the remainder of the
> > > arguments are a list of which local.config.* files to use.
> > > That way we can run a test against different servers and also
> > > different configurations.
> >
> > This is smart. :)
> >
> > > I try to keep the target names meaningful, so
> > >
> > > "smb3" just means a basic smb3   windows 2016 server running in a vm
> > > "smb3samba" is just a generic smb3 server with samba.
> > > etc etc.
> > >
> > >
> > > We started doing this upstream about a year ago which is when I
> > > started setting it up
> > > and now I am honest but since we do this for every pull request we
> > > send to linus,
> > > we have SIGNIFICANTLY increased the quality of cifs.ko.
> >
> > /Thumbs up!
> >
> > >
> > > We run quite a lot of tests nowadays and I always try to remind Steve
> > > to "please put a link to the buildbot result in the pull request"
> > > to bring awareness to our buildbot.
> > >
> > > For upstream, here is the most recent test:
> > >
> > http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/builders/2/builds/249
> > >
> > > cifs-testing is the most generic test target. We also have specific
> > > targets that are just for Azure, and other targets, but the aim is
> > > that the canonical cifs-target tests will all pass before we pass a
> > > pull request onto linus.
> > >
> > >
> > > Please contact me directly if you want more info about the buildbot. I
> > > have one setup internally at rh too that runs many, but not all, the
> > > tests we run upstream.
> >
> > Sure! Have replied to you. Thank you very much for sharing!
> >
> > If anybody resting cifs or I know that the expected setup and result of
> > every tests for you developers, it will save much time of investigation
> > and report false alarms.
> >
> > After several months working on cifs, I've got a feeling that cifs is
> > much different from other Linux filesystems. Maybe because of Windows
> > or the SMB protocols. I'm wondering that xfstests which is a typical
> > testsuite for Unix/Linux filesystems, maybe is not that suitable for
> > cifs. I agree with Steve that there are hundreds of tests suitable.
> > If, only if, :) if we could know how MS testing NTFS or SMB protocols,
> > would that help or more suitable?
> >
> >
> > Thanks!
> > Murphy
> >
> >
> > > Any help to maintain and expand the tests are super welcome.
> > >
> > >
> > > Rant-off
> > > ronnie sahlberg
> > >
> > > >
> > > > >
> > > > > But would really like to add test targets for other common servers
> > and
> > > > > in a perfect world would like to be able to have external test
> > > > > automation pull our trees periodically to run these (similar to what
> > > > > is done for the six targets above):
> > > > >   - Macs
> > > > >   - NetApp
> > > > > (and any others of interest to the community)
> > > > >
> > > > > Obviously xfstest has a few hundred tests which only make sense for
> > > > > local file systems with block devices, but there are hundreds of
> > > > > xfstests of value, and most should run on cifs.ko and we are working
> > > > > through them but already have a very good set running.
> > > >
> > > > Agree!  Thanks!
> > > >
> > > > Murphy
> > > >
> > > > >
> > > > > On Mon, Sep 9, 2019 at 5:50 PM Aurélien Aptel <aaptel@suse.com>
> > wrote:
> > > > > >
> > > > > > "Murphy Zhou" <jencce.kernel@gmail.com> writes:
> > > > > > > As $subject. Is this wiki being maintained ?
> > > > > > >
> > > > > > > Looks like the last update was in January 2019.
> > > > > > >
> > > > > > > https://wiki.samba.org/index.php/Xfstesting-cifs#Exclusion_files
> > > > > >
> > > > > > We have a buildbot running xfstests relatively often here
> > > > > >
> > > > > > http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/
> > > > > >
> > > > > > Each group has slightly different tests run, you can see which one
> > gets
> > > > > > run by clicking on a build:
> > > > > >
> > > > > >
> > http://smb3-test-rhel-75.southcentralus.cloudapp.azure.com/#/builders/2/builds/249
> > > > > >
> > > > > > Overall xfstests+cifs is very finicky and frustrating to get
> > working
> > > > > > reliably. Not to mention long. So good luck :)
> > > > > >
> > > > > > Cheers,
> > > > > > --
> > > > > > Aurélien Aptel / SUSE Labs Samba Team
> > > > > > GPG: 1839 CB5F 9F5B FB9B AA97  8C99 03C8 A49B 521B D5D3
> > > > > > SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409
> > Nürnberg, DE
> > > > > > GF: Felix Imendörffer, Mary Higgins, Sri Rasiah HRB 247165 (AG
> > München)
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Thanks,
> > > > >
> > > > > Steve
> >

  parent reply	other threads:[~2019-09-14  0:45 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-09 10:41 Are the xfstests exclusion files on wiki.samba.org up to date? Murphy Zhou
2019-09-09 10:50 ` Aurélien Aptel
2019-09-09 23:23   ` Steve French
2019-09-10 10:39     ` ronnie sahlberg
2019-09-10 11:41       ` Aurélien Aptel
2019-09-12  8:50         ` Murphy Zhou
2019-09-12  8:49     ` Murphy Zhou
2019-09-12  9:06       ` ronnie sahlberg
2019-09-14  0:02         ` Murphy Zhou
     [not found]           ` <CAH2r5mv5RqAfAMb2oWa+10bqUa+6kUMO3HKPG1s2z0PRqOeKEg@mail.gmail.com>
2019-09-14  0:44             ` Murphy Zhou [this message]
2019-09-12  8:45   ` Murphy Zhou

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=20190914004451.k2fhirxo54hgh5ln@XZHOUW.usersys.redhat.com \
    --to=jencce.kernel@gmail.com \
    --cc=aaptel@suse.com \
    --cc=linux-cifs@vger.kernel.org \
    --cc=ronniesahlberg@gmail.com \
    --cc=smfrench@gmail.com \
    /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).