| Message ID | 20120711203151.3767.79006.stgit@degas.1015granger.net (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, 2012-07-11 at 16:31 -0400, Chuck Lever wrote: > If the state manager thread is not actually able to fully recover from > some situation, it wakes up waiters, who kick off a new state manager > thread. Quite often the fresh invocation of the state manager is just > as successful. > > This results in a livelock as the client dumps thousands of NFS > requests a second on the network in a vain attempt to recover. Not > very friendly. > > To mitigate this situation, add a delay in the state manager after > an unhandled error, so that the client sends just a few requests > every second in this case. I assume that this was intended to have a s-o-b line? > --- > > fs/nfs/nfs4state.c | 1 + > 1 files changed, 1 insertions(+), 0 deletions(-) > > diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c > index 5e3bebc..38959eb 100644 > --- a/fs/nfs/nfs4state.c > +++ b/fs/nfs/nfs4state.c > @@ -2151,6 +2151,7 @@ static void nfs4_state_manager(struct nfs_client *clp) > out_error: > pr_warn_ratelimited("NFS: state manager failed on NFSv4 server %s" > " with error %d\n", clp->cl_hostname, -status); > + ssleep(1); > nfs4_end_drain_session(clp); > nfs4_clear_state_manager_bit(clp); > } > -- Trond Myklebust Linux NFS client maintainer NetApp Trond.Myklebust@netapp.com www.netapp.com
On Jul 17, 2012, at 2:06 PM, Myklebust, Trond wrote: > On Wed, 2012-07-11 at 16:31 -0400, Chuck Lever wrote: >> If the state manager thread is not actually able to fully recover from >> some situation, it wakes up waiters, who kick off a new state manager >> thread. Quite often the fresh invocation of the state manager is just >> as successful. >> >> This results in a livelock as the client dumps thousands of NFS >> requests a second on the network in a vain attempt to recover. Not >> very friendly. >> >> To mitigate this situation, add a delay in the state manager after >> an unhandled error, so that the client sends just a few requests >> every second in this case. > > > I assume that this was intended to have a s-o-b line? Yes. Actually it was submitted mainly for discussion, which is why the S-O-B was accidentally omitted. But if you feel it is appropriate for mainline now, I'm happy to see it merged. > >> --- >> >> fs/nfs/nfs4state.c | 1 + >> 1 files changed, 1 insertions(+), 0 deletions(-) >> >> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c >> index 5e3bebc..38959eb 100644 >> --- a/fs/nfs/nfs4state.c >> +++ b/fs/nfs/nfs4state.c >> @@ -2151,6 +2151,7 @@ static void nfs4_state_manager(struct nfs_client *clp) >> out_error: >> pr_warn_ratelimited("NFS: state manager failed on NFSv4 server %s" >> " with error %d\n", clp->cl_hostname, -status); >> + ssleep(1); >> nfs4_end_drain_session(clp); >> nfs4_clear_state_manager_bit(clp); >> } >> > > -- > Trond Myklebust > Linux NFS client maintainer > > NetApp > Trond.Myklebust@netapp.com > www.netapp.com >
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 5e3bebc..38959eb 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -2151,6 +2151,7 @@ static void nfs4_state_manager(struct nfs_client *clp) out_error: pr_warn_ratelimited("NFS: state manager failed on NFSv4 server %s" " with error %d\n", clp->cl_hostname, -status); + ssleep(1); nfs4_end_drain_session(clp); nfs4_clear_state_manager_bit(clp); }