Re: [PATCH] ceph: show max caps in debugfs caps file

From: Yanhu Cao
Date: Thu May 21 2020 - 08:21:13 EST


On Thu, May 21, 2020 at 7:09 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
>
> On Thu, 2020-05-21 at 17:38 +0800, Yanhu Cao wrote:
> > before
> > ------
> > total 1026
> > avail 1024
> > used 2
> > reserved 0
> > min 1024
> >
> > after
> > ------
> > total 1026
> > avail 1024
> > used 2
> > max 2048
> > reserved 0
> > min 1024
> >
> > Signed-off-by: Yanhu Cao <gmayyyha@xxxxxxxxx>
> > ---
> > fs/ceph/caps.c | 6 ++++--
> > fs/ceph/debugfs.c | 7 ++++---
> > fs/ceph/super.h | 2 +-
> > 3 files changed, 9 insertions(+), 6 deletions(-)
> >
> > diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
> > index 5f3aa4d607de..e2c759a2ef35 100644
> > --- a/fs/ceph/caps.c
> > +++ b/fs/ceph/caps.c
> > @@ -404,8 +404,8 @@ void ceph_put_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap)
> > }
> >
> > void ceph_reservation_status(struct ceph_fs_client *fsc,
> > - int *total, int *avail, int *used, int *reserved,
> > - int *min)
> > + int *total, int *avail, int *used, int *max,
> > + int *reserved, int *min)
> > {
> > struct ceph_mds_client *mdsc = fsc->mdsc;
> >
> > @@ -417,6 +417,8 @@ void ceph_reservation_status(struct ceph_fs_client *fsc,
> > *avail = mdsc->caps_avail_count;
> > if (used)
> > *used = mdsc->caps_use_count;
> > + if (max)
> > + *max = mdsc->caps_use_max;
>
> Can you lay out what value this will provide? I'm not convinced that
> this information is really that helpful:
>
> mdsc->caps_use_max is just set to the value of the "caps_max" mount
> option, and that information is displayed in /proc/mounts if it's not
> set to the default.
>
> What might be more interesting is to track the most recent "max_caps"
> value sent by the MDS (see the CEPH_SESSION_RECALL_STATE message
> handling). Tracking that would give us a more dynamic view of the
> current maximum requested by the MDS, which is often going to be less
> than what "caps_max" was set to at mount time.

Do you mean the 'mds_recall_max_caps'? which can be set by the MDS.
Clients use this value every time to trim caps.

There is an option mds_max_caps_per_client which is a soft limit,
which is determined by the behavior of the client.
and we recently encounter a warning '1 MDSs report oversized
cache'(ceph-v12.2.12: mds_cache_memory_limit=64G, used=100G),
Therefore, the effect is not good.

So we want to know whether the caps held by the client exceed caps_max
through the debugfs caps file (default or mount option).


>
> > if (reserved)
> > *reserved = mdsc->caps_reserve_count;
> > if (min)
> > diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c
> > index 481ac97b4d25..942004376588 100644
> > --- a/fs/ceph/debugfs.c
> > +++ b/fs/ceph/debugfs.c
> > @@ -138,16 +138,17 @@ static int caps_show(struct seq_file *s, void *p)
> > {
> > struct ceph_fs_client *fsc = s->private;
> > struct ceph_mds_client *mdsc = fsc->mdsc;
> > - int total, avail, used, reserved, min, i;
> > + int total, avail, used, max, reserved, min, i;
> > struct cap_wait *cw;
> >
> > - ceph_reservation_status(fsc, &total, &avail, &used, &reserved, &min);
> > + ceph_reservation_status(fsc, &total, &avail, &used, &max,
> > + &reserved, &min);
> > seq_printf(s, "total\t\t%d\n"
> > "avail\t\t%d\n"
> > "used\t\t%d\n"
> > "reserved\t%d\n"
> > "min\t\t%d\n\n",
> > - total, avail, used, reserved, min);
> > + total, avail, used, max, reserved, min);
> > seq_printf(s, "ino issued implemented\n");
> > seq_printf(s, "-----------------------------------------------\n");
> >
> > diff --git a/fs/ceph/super.h b/fs/ceph/super.h
> > index 60aac3aee055..79aa42d9336c 100644
> > --- a/fs/ceph/super.h
> > +++ b/fs/ceph/super.h
> > @@ -700,7 +700,7 @@ extern void ceph_unreserve_caps(struct ceph_mds_client *mdsc,
> > struct ceph_cap_reservation *ctx);
> > extern void ceph_reservation_status(struct ceph_fs_client *client,
> > int *total, int *avail, int *used,
> > - int *reserved, int *min);
> > + int *max, int *reserved, int *min);
> >
> >
> >
>
> --
> Jeff Layton <jlayton@xxxxxxxxxx>
>