diff mbox series

[v2] ceph: use frag's MDS in either mode

Message ID 20200731082513.11806-1-gmayyyha@gmail.com (mailing list archive)
State New, archived
Headers show
Series [v2] ceph: use frag's MDS in either mode | expand

Commit Message

Yanhu Cao July 31, 2020, 8:25 a.m. UTC
When doing some tests with multiple mds, there are many
mds forwarding requests between them and then the client request
is resent.

If the request is a modification operation and the mode is set to USE_AUTH_MDS.
Auth mds should be selected to handle the request, and if auth mds for
frag is already set, it can be returned directly without the logic behind it.

But the current logic doesn't return directly because the condition
'mode == USE_AUTH_MDS', and sometimes frag's mds is not equal to
cap's session mds, which then causes the request to be resent.

Signed-off-by: Yanhu Cao <gmayyyha@gmail.com>
---
 fs/ceph/mds_client.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

Comments

Jeff Layton July 31, 2020, 11:57 a.m. UTC | #1
On Fri, 2020-07-31 at 16:25 +0800, Yanhu Cao wrote:
> When doing some tests with multiple mds, there are many
> mds forwarding requests between them and then the client request
> is resent.
> 
> If the request is a modification operation and the mode is set to USE_AUTH_MDS.
> Auth mds should be selected to handle the request, and if auth mds for
> frag is already set, it can be returned directly without the logic behind it.
> 
> But the current logic doesn't return directly because the condition
> 'mode == USE_AUTH_MDS', and sometimes frag's mds is not equal to
> cap's session mds, which then causes the request to be resent.
> 
> Signed-off-by: Yanhu Cao <gmayyyha@gmail.com>
> ---
>  fs/ceph/mds_client.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
> index a50497142e59..b2255a9be7c0 100644
> --- a/fs/ceph/mds_client.c
> +++ b/fs/ceph/mds_client.c
> @@ -1103,8 +1103,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc,
>  				     frag.frag, mds);
>  				if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
>  				    CEPH_MDS_STATE_ACTIVE) {
> -					if (mode == USE_ANY_MDS &&
> -					    !ceph_mdsmap_is_laggy(mdsc->mdsmap,
> +					if (!ceph_mdsmap_is_laggy(mdsc->mdsmap,
>  								  mds))
>  						goto out;
>  				}


Much better! A good changelog is as important as the patch, esp in code
like this that is quite non-obvious.

Merged into testing branch. I did clean up the changelog a bit before I
merged it. Please take a look when you have time and let me know if you
think I got it wrong.

Thanks!
Yanhu Cao Aug. 3, 2020, 1:25 a.m. UTC | #2
On Fri, Jul 31, 2020 at 7:57 PM Jeff Layton <jlayton@kernel.org> wrote:
>
> On Fri, 2020-07-31 at 16:25 +0800, Yanhu Cao wrote:
> > When doing some tests with multiple mds, there are many
> > mds forwarding requests between them and then the client request
> > is resent.
> >
> > If the request is a modification operation and the mode is set to USE_AUTH_MDS.
> > Auth mds should be selected to handle the request, and if auth mds for
> > frag is already set, it can be returned directly without the logic behind it.
> >
> > But the current logic doesn't return directly because the condition
> > 'mode == USE_AUTH_MDS', and sometimes frag's mds is not equal to
> > cap's session mds, which then causes the request to be resent.
> >
> > Signed-off-by: Yanhu Cao <gmayyyha@gmail.com>
> > ---
> >  fs/ceph/mds_client.c | 3 +--
> >  1 file changed, 1 insertion(+), 2 deletions(-)
> >
> > diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
> > index a50497142e59..b2255a9be7c0 100644
> > --- a/fs/ceph/mds_client.c
> > +++ b/fs/ceph/mds_client.c
> > @@ -1103,8 +1103,7 @@ static int __choose_mds(struct ceph_mds_client *mdsc,
> >                                    frag.frag, mds);
> >                               if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
> >                                   CEPH_MDS_STATE_ACTIVE) {
> > -                                     if (mode == USE_ANY_MDS &&
> > -                                         !ceph_mdsmap_is_laggy(mdsc->mdsmap,
> > +                                     if (!ceph_mdsmap_is_laggy(mdsc->mdsmap,
> >                                                                 mds))
> >                                               goto out;
> >                               }
>
>
> Much better! A good changelog is as important as the patch, esp in code
> like this that is quite non-obvious.
>
> Merged into testing branch. I did clean up the changelog a bit before I
> merged it. Please take a look when you have time and let me know if you
> think I got it wrong.
LGTM.

Thanks.
>
> Thanks!
> --
> Jeff Layton <jlayton@kernel.org>
>
diff mbox series

Patch

diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index a50497142e59..b2255a9be7c0 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1103,8 +1103,7 @@  static int __choose_mds(struct ceph_mds_client *mdsc,
 				     frag.frag, mds);
 				if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
 				    CEPH_MDS_STATE_ACTIVE) {
-					if (mode == USE_ANY_MDS &&
-					    !ceph_mdsmap_is_laggy(mdsc->mdsmap,
+					if (!ceph_mdsmap_is_laggy(mdsc->mdsmap,
 								  mds))
 						goto out;
 				}