2011-11-18 16:20:38

by John Hughes

[permalink] [raw]
Subject: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

Don't write anything to the log if a user Kerberos ticket expires or is missing,
these are not kernel problems.

Signed-off-by: John Hughes<[email protected]>

diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 39914be..18897cb 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -1339,14 +1339,6 @@ static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
}

-static void nfs4_warn_keyexpired(const char *s)
-{
- printk_ratelimited(KERN_WARNING "Error: state manager"
- " encountered RPCSEC_GSS session"
- " expired against NFSv4 server %s.\n",
- s);
-}
-
static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
{
switch (error) {
@@ -1378,7 +1370,6 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
return 0;
case -EKEYEXPIRED:
/* Nothing we can do */
- nfs4_warn_keyexpired(clp->cl_hostname);
return 0;
}
return error;
@@ -1620,7 +1611,6 @@ static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
break;

case -EKEYEXPIRED:
- nfs4_warn_keyexpired(clp->cl_hostname);
case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
* in nfs4_exchange_id */
default:
@@ -1728,8 +1718,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
} while (atomic_read(&clp->cl_count)> 1);
return;
out_error:
- printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
- " with error %d\n", clp->cl_hostname, -status);
+ if (status != -EACCESS)
+ printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
+ " with error %d\n", clp->cl_hostname, -status);
nfs4_end_drain_session(clp);
nfs4_clear_state_manager_bit(clp);
}




2011-11-18 17:51:17

by peter.staubach

[permalink] [raw]
Subject: RE: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

The log messages are for the poor admin who has to figure out why nothing is working for the user.

ps


-----Original Message-----
From: John Hughes [mailto:[email protected]]
Sent: Friday, November 18, 2011 11:52 AM
To: Staubach, Peter
Cc: [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

On 18/11/11 17:27, [email protected] wrote:
> Given that the poor user will be sitting, with almost nothing working because of the expired ticket, how does the user figure out what is going on or how does the admin help him?
>
> Some guidance seems like a good thing. Is there really a problem with these messages being printed to the log?
If you're sat at your console with nothing responding how are you going
to see the log messages?

The "ticket expired" messages are rate limited.

The "Error: state manager failed on NFSv4 server" messages are not rate
limited, but if you use my "rpc.gssd -e" hack they'll be written to the
log every 5 seconds all night long.



2011-11-18 18:30:16

by Myklebust, Trond

[permalink] [raw]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

On Fri, 2011-11-18 at 17:20 +0100, John Hughes wrote:
> Don't write anything to the log if a user Kerberos ticket expires or is missing,
> these are not kernel problems.
>
> Signed-off-by: John Hughes<[email protected]>
>
> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
> index 39914be..18897cb 100644
> --- a/fs/nfs/nfs4state.c
> +++ b/fs/nfs/nfs4state.c
> @@ -1339,14 +1339,6 @@ static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
> nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
> }
>
> -static void nfs4_warn_keyexpired(const char *s)
> -{
> - printk_ratelimited(KERN_WARNING "Error: state manager"
> - " encountered RPCSEC_GSS session"
> - " expired against NFSv4 server %s.\n",
> - s);
> -}

NACK. This is useful.

> -
> static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
> {
> switch (error) {
> @@ -1378,7 +1370,6 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
> return 0;
> case -EKEYEXPIRED:
> /* Nothing we can do */
> - nfs4_warn_keyexpired(clp->cl_hostname);
> return 0;
> }
> return error;
> @@ -1620,7 +1611,6 @@ static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
> break;
>
> case -EKEYEXPIRED:
> - nfs4_warn_keyexpired(clp->cl_hostname);

NACK. This means that the server can't even recover it's lease. It
shouldn't even trigger if you install a machine credential.

> case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
> * in nfs4_exchange_id */
> default:
> @@ -1728,8 +1718,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
> } while (atomic_read(&clp->cl_count)> 1);
> return;
> out_error:
> - printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
> - " with error %d\n", clp->cl_hostname, -status);
> + if (status != -EACCESS)
> + printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
> + " with error %d\n", clp->cl_hostname, -status);

Obviously not even compile tested...

> nfs4_end_drain_session(clp);
> nfs4_clear_state_manager_bit(clp);
> }
>
>

--
Trond Myklebust
Linux NFS client maintainer

NetApp
[email protected]
http://www.netapp.com


2011-11-18 19:17:26

by John Hughes

[permalink] [raw]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

On 11/18/2011 07:30 PM, Trond Myklebust wrote:
> On Fri, 2011-11-18 at 17:20 +0100, John Hughes wrote:
>
>> Don't write anything to the log if a user Kerberos ticket expires or is missing,
>> these are not kernel problems.
>>
>> Signed-off-by: John Hughes<[email protected]>
>>
>> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
>> index 39914be..18897cb 100644
>> --- a/fs/nfs/nfs4state.c
>> +++ b/fs/nfs/nfs4state.c
>> @@ -1339,14 +1339,6 @@ static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
>> nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
>> }
>>
>> -static void nfs4_warn_keyexpired(const char *s)
>> -{
>> - printk_ratelimited(KERN_WARNING "Error: state manager"
>> - " encountered RPCSEC_GSS session"
>> - " expired against NFSv4 server %s.\n",
>> - s);
>> -}
>>
> NACK. This is useful.
>
>
>> -
>> static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
>> {
>> switch (error) {
>> @@ -1378,7 +1370,6 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
>> return 0;
>> case -EKEYEXPIRED:
>> /* Nothing we can do */
>> - nfs4_warn_keyexpired(clp->cl_hostname);
>> return 0;
>> }
>> return error;
>>


>> @@ -1620,7 +1611,6 @@ static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
>> break;
>>
>> case -EKEYEXPIRED:
>> - nfs4_warn_keyexpired(clp->cl_hostname);
>>
> NACK. This means that the server can't even recover it's lease. It
> shouldn't even trigger if you install a machine credential.
>

Ok, couldn't see why there were two such warnings. Maybe the error
should be more alarming in this case?

>
>> case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
>> * in nfs4_exchange_id */
>> default:
>> @@ -1728,8 +1718,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
>> } while (atomic_read(&clp->cl_count)> 1);
>> return;
>> out_error:
>> - printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
>> - " with error %d\n", clp->cl_hostname, -status);
>> + if (status != -EACCESS)
>> + printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
>> + " with error %d\n", clp->cl_hostname, -status);
>>
> Obviously not even compile tested...
>

Oh dear, my sins have found me out. Sorry.

2011-11-18 16:52:00

by John Hughes

[permalink] [raw]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

On 18/11/11 17:27, [email protected] wrote:
> Given that the poor user will be sitting, with almost nothing working because of the expired ticket, how does the user figure out what is going on or how does the admin help him?
>
> Some guidance seems like a good thing. Is there really a problem with these messages being printed to the log?
If you're sat at your console with nothing responding how are you going
to see the log messages?

The "ticket expired" messages are rate limited.

The "Error: state manager failed on NFSv4 server" messages are not rate
limited, but if you use my "rpc.gssd -e" hack they'll be written to the
log every 5 seconds all night long.


2011-11-18 16:58:54

by Matt W. Benjamin

[permalink] [raw]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

Please forgive me for speaking this aloud, but, dbus?

Matt

----- "John Hughes" <[email protected]> wrote:

> On 18/11/11 17:27, [email protected] wrote:
> > Given that the poor user will be sitting, with almost nothing
> working because of the expired ticket, how does the user figure out
> what is going on or how does the admin help him?
> >
> > Some guidance seems like a good thing. Is there really a problem
> with these messages being printed to the log?
> If you're sat at your console with nothing responding how are you
> going
> to see the log messages?
>
> The "ticket expired" messages are rate limited.
>
> The "Error: state manager failed on NFSv4 server" messages are not
> rate
> limited, but if you use my "rpc.gssd -e" hack they'll be written to
> the
> log every 5 seconds all night long.
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs"
> in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html

--

Matt Benjamin

The Linux Box
206 South Fifth Ave. Suite 150
Ann Arbor, MI 48104

http://linuxbox.com

tel. 734-761-4689
fax. 734-769-8938
cel. 734-216-5309

2011-11-18 16:28:01

by peter.staubach

[permalink] [raw]
Subject: RE: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

Given that the poor user will be sitting, with almost nothing working because of the expired ticket, how does the user figure out what is going on or how does the admin help him?

Some guidance seems like a good thing. Is there really a problem with these messages being printed to the log?

Thanx...

ps


-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of John Hughes
Sent: Friday, November 18, 2011 11:21 AM
To: Trond Myklebust
Cc: [email protected]; [email protected]
Subject: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

Don't write anything to the log if a user Kerberos ticket expires or is missing,
these are not kernel problems.

Signed-off-by: John Hughes<[email protected]>

diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 39914be..18897cb 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -1339,14 +1339,6 @@ static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
}

-static void nfs4_warn_keyexpired(const char *s)
-{
- printk_ratelimited(KERN_WARNING "Error: state manager"
- " encountered RPCSEC_GSS session"
- " expired against NFSv4 server %s.\n",
- s);
-}
-
static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
{
switch (error) {
@@ -1378,7 +1370,6 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
return 0;
case -EKEYEXPIRED:
/* Nothing we can do */
- nfs4_warn_keyexpired(clp->cl_hostname);
return 0;
}
return error;
@@ -1620,7 +1611,6 @@ static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
break;

case -EKEYEXPIRED:
- nfs4_warn_keyexpired(clp->cl_hostname);
case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
* in nfs4_exchange_id */
default:
@@ -1728,8 +1718,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
} while (atomic_read(&clp->cl_count)> 1);
return;
out_error:
- printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
- " with error %d\n", clp->cl_hostname, -status);
+ if (status != -EACCESS)
+ printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
+ " with error %d\n", clp->cl_hostname, -status);
nfs4_end_drain_session(clp);
nfs4_clear_state_manager_bit(clp);
}

2011-11-18 16:34:38

by Anna Schumaker

[permalink] [raw]
Subject: Re: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors

On Fri 18 Nov 2011 11:27:37 AM EST, [email protected] wrote:
> Given that the poor user will be sitting, with almost nothing working because of the expired ticket, how does the user figure out what is going on or how does the admin help him?
>
> Some guidance seems like a good thing. Is there really a problem with these messages being printed to the log?

I agree. Rather than removing the messages, is there a way to slow
them down even further so the logs don't fill up as quickly?

>
> Thanx...
>
> ps
>
>
> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of John Hughes
> Sent: Friday, November 18, 2011 11:21 AM
> To: Trond Myklebust
> Cc: [email protected]; [email protected]
> Subject: [PATCH] Don't spam kernel log with EKEYEXPIRED and EACCESS errors
>
> Don't write anything to the log if a user Kerberos ticket expires or is missing,
> these are not kernel problems.
>
> Signed-off-by: John Hughes<[email protected]>
>
> diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
> index 39914be..18897cb 100644
> --- a/fs/nfs/nfs4state.c
> +++ b/fs/nfs/nfs4state.c
> @@ -1339,14 +1339,6 @@ static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
> nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
> }
>
> -static void nfs4_warn_keyexpired(const char *s)
> -{
> - printk_ratelimited(KERN_WARNING "Error: state manager"
> - " encountered RPCSEC_GSS session"
> - " expired against NFSv4 server %s.\n",
> - s);
> -}
> -
> static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
> {
> switch (error) {
> @@ -1378,7 +1370,6 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error)
> return 0;
> case -EKEYEXPIRED:
> /* Nothing we can do */
> - nfs4_warn_keyexpired(clp->cl_hostname);
> return 0;
> }
> return error;
> @@ -1620,7 +1611,6 @@ static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
> break;
>
> case -EKEYEXPIRED:
> - nfs4_warn_keyexpired(clp->cl_hostname);
> case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
> * in nfs4_exchange_id */
> default:
> @@ -1728,8 +1718,9 @@ static void nfs4_state_manager(struct nfs_client *clp)
> } while (atomic_read(&clp->cl_count)> 1);
> return;
> out_error:
> - printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
> - " with error %d\n", clp->cl_hostname, -status);
> + if (status != -EACCESS)
> + printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
> + " with error %d\n", clp->cl_hostname, -status);
> nfs4_end_drain_session(clp);
> nfs4_clear_state_manager_bit(clp);
> }
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html