[PATCH 0/2] nfs-utils: detect and log write errors

[PATCH 0/2] nfs-utils: detect and log write errors

[Kevin Coffman]

Address at least one cause of a problem that has been reported
where it appears that a gss context was successfully created,
but the server returns GSS_S_NO_CONTEXT.

---

Kevin Coffman (2):
      svcgssd: check the return code from qword_eol() and log failures
      cacheio: return any original error from qword_eol


 support/nfs/cacheio.c     |   10 ++++++++--
 utils/gssd/svcgssd_proc.c |    4 ++++
 2 files changed, 12 insertions(+), 2 deletions(-)

[PATCH 1/2] cacheio: return any original error from qword_eol

[Kevin Coffman]

From: Kevin Coffman <kwc@citi.umich.edu>

If the initial fflush() fails in qword_eol, log the failure
and return the indication of the original failure, not the
successful cover-up.

Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
---

 support/nfs/cacheio.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)


diff --git a/support/nfs/cacheio.c b/support/nfs/cacheio.c
index f303734..6a6ed5a 100644
--- a/support/nfs/cacheio.c
+++ b/support/nfs/cacheio.c
@@ -154,15 +154,21 @@ int qword_eol(FILE *f)
 
 	fprintf(f,"\n");
 	err = fflush(f);
+	if (err) {
+		xlog_warn("qword_eol: fflush failed: errno %d (%s)",
+			    errno, strerror(errno));
+	}
 	/*
 	 * We must send one line (and one line only) in a single write
 	 * call.  In case of a write error, libc may accumulate the
 	 * unwritten data and try to write it again later, resulting in a
 	 * multi-line write.  So we must explicitly ask it to throw away
-	 * any such cached data:
+	 * any such cached data.  But we return any original error
+	 * indication to the caller.
 	 */
 	__fpurge(f);
-	return fflush(f);
+	fflush(f);
+	return err;
 }
 
 

[PATCH 2/2] svcgssd: check the return code from qword_eol() and log failures

[Kevin Coffman]

From: Kevin Coffman <kwc@citi.umich.edu>

If qword_eol() fails while writing the context information, log
an indication of the failure.

This addresses at least one cause of the intermittent, and
previously undiagnosed, problem of the server returning
GSS_S_NO_CONTEXT when a context was seemingly successfully
created and sent down to the kernel.  In my case there was a
mis-match between kernel and user-land configuration resulting in
the proper kernel module not being loaded. Therefore the write
of the context failed, but was not logged by svcgssd.  When the
kernel goes to find the resulting context, it was really not
there and correctly returned GSS_S_NO_CONTEXT to the client.

Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
---

 utils/gssd/svcgssd_proc.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)


diff --git a/utils/gssd/svcgssd_proc.c b/utils/gssd/svcgssd_proc.c
index b390bea..6f2ba61 100644
--- a/utils/gssd/svcgssd_proc.c
+++ b/utils/gssd/svcgssd_proc.c
@@ -109,6 +109,10 @@ do_svc_downcall(gss_buffer_desc *out_handle, struct svc_cred *cred,
 	qword_print(f, fname);
 	qword_printhex(f, context_token->value, context_token->length);
 	err = qword_eol(f);
+	if (err) {
+		printerr(1, "WARNING: error writing to downcall channel "
+			 "%s: %s\n", SVCGSSD_CONTEXT_CHANNEL, strerror(errno));
+	}
 	fclose(f);
 	return err;
 out_err:

Re: [PATCH 0/2] nfs-utils: detect and log write errors

[Steve Dickson]

Kevin Coffman wrote:
> Address at least one cause of a problem that has been reported
> where it appears that a gss context was successfully created,
> but the server returns GSS_S_NO_CONTEXT.
> 
> ---
> 
> Kevin Coffman (2):
>       svcgssd: check the return code from qword_eol() and log failures
>       cacheio: return any original error from qword_eol
> 
> 
>  support/nfs/cacheio.c     |   10 ++++++++--
>  utils/gssd/svcgssd_proc.c |    4 ++++
>  2 files changed, 12 insertions(+), 2 deletions(-)
Both patches committed...

steved.