[PATCH 2/4] target: log Data-Out timeouts as errors

[PATCH 2/4] target: log Data-Out timeouts as errors

[David Disseldorp]

Data-Out timeouts resulting in connection outages should be logged as
errors.

Signed-off-by: David Disseldorp <ddiss@suse.de>
---
 drivers/target/iscsi/iscsi_target_erl1.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/target/iscsi/iscsi_target_erl1.c b/drivers/target/iscsi/iscsi_target_erl1.c
index 5efa42b939a1..d75cc11ca83c 100644
--- a/drivers/target/iscsi/iscsi_target_erl1.c
+++ b/drivers/target/iscsi/iscsi_target_erl1.c
@@ -1169,13 +1169,13 @@ void iscsit_handle_dataout_timeout(struct timer_list *t)
 	na = iscsit_tpg_get_node_attrib(sess);
 
 	if (!sess->sess_ops->ErrorRecoveryLevel) {
-		pr_debug("Unable to recover from DataOut timeout while"
-			" in ERL=0.\n");
+		pr_err("Unable to recover from DataOut timeout while"
+			" in ERL=0, closing iSCSI connection.\n");
 		goto failure;
 	}
 
 	if (++cmd->dataout_timeout_retries = na->dataout_timeout_retries) {
-		pr_debug("Command ITT: 0x%08x exceeded max retries"
+		pr_err("Command ITT: 0x%08x exceeded max retries"
 			" for DataOUT timeout %u, closing iSCSI connection.\n",
 			cmd->init_task_tag, na->dataout_timeout_retries);
 		goto failure;
-- 
2.13.7

Re: [PATCH 2/4] target: log Data-Out timeouts as errors

[Mike Christie]

On 10/09/2018 06:37 AM, David Disseldorp wrote:
> Data-Out timeouts resulting in connection outages should be logged as
> errors.
> 
> Signed-off-by: David Disseldorp <ddiss@suse.de>
> ---
>  drivers/target/iscsi/iscsi_target_erl1.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/target/iscsi/iscsi_target_erl1.c b/drivers/target/iscsi/iscsi_target_erl1.c
> index 5efa42b939a1..d75cc11ca83c 100644
> --- a/drivers/target/iscsi/iscsi_target_erl1.c
> +++ b/drivers/target/iscsi/iscsi_target_erl1.c
> @@ -1169,13 +1169,13 @@ void iscsit_handle_dataout_timeout(struct timer_list *t)
>  	na = iscsit_tpg_get_node_attrib(sess);
>  
>  	if (!sess->sess_ops->ErrorRecoveryLevel) {
> -		pr_debug("Unable to recover from DataOut timeout while"
> -			" in ERL=0.\n");
> +		pr_err("Unable to recover from DataOut timeout while"
> +			" in ERL=0, closing iSCSI connection.\n");
>  		goto failure;
>  	}
>  
>  	if (++cmd->dataout_timeout_retries = na->dataout_timeout_retries) {
> -		pr_debug("Command ITT: 0x%08x exceeded max retries"
> +		pr_err("Command ITT: 0x%08x exceeded max retries"
>  			" for DataOUT timeout %u, closing iSCSI connection.\n",
>  			cmd->init_task_tag, na->dataout_timeout_retries);
>  		goto failure;
> 

For this and the nop one I think we should print something so we know
what session failed. For the nop one the sid can change on relogin if
this was the only session, so you can't later match it from the logs and
configfs.

Maybe in the nop/dataout failure path add the I_T nexus values for the
failed session in the failure message or maybe for both messages log the
{sid, cid}  but then on login also log a message indicating a connection
logged in and the I_T nexus to sid/cid mapping.

Re: [PATCH 2/4] target: log Data-Out timeouts as errors

[Mike Christie]

On 10/09/2018 10:56 AM, Mike Christie wrote:
> On 10/09/2018 06:37 AM, David Disseldorp wrote:
>> Data-Out timeouts resulting in connection outages should be logged as
>> errors.
>>
>> Signed-off-by: David Disseldorp <ddiss@suse.de>
>> ---
>>  drivers/target/iscsi/iscsi_target_erl1.c | 6 +++---
>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/target/iscsi/iscsi_target_erl1.c b/drivers/target/iscsi/iscsi_target_erl1.c
>> index 5efa42b939a1..d75cc11ca83c 100644
>> --- a/drivers/target/iscsi/iscsi_target_erl1.c
>> +++ b/drivers/target/iscsi/iscsi_target_erl1.c
>> @@ -1169,13 +1169,13 @@ void iscsit_handle_dataout_timeout(struct timer_list *t)
>>  	na = iscsit_tpg_get_node_attrib(sess);
>>  
>>  	if (!sess->sess_ops->ErrorRecoveryLevel) {
>> -		pr_debug("Unable to recover from DataOut timeout while"
>> -			" in ERL=0.\n");
>> +		pr_err("Unable to recover from DataOut timeout while"
>> +			" in ERL=0, closing iSCSI connection.\n");
>>  		goto failure;
>>  	}
>>  
>>  	if (++cmd->dataout_timeout_retries = na->dataout_timeout_retries) {
>> -		pr_debug("Command ITT: 0x%08x exceeded max retries"
>> +		pr_err("Command ITT: 0x%08x exceeded max retries"
>>  			" for DataOUT timeout %u, closing iSCSI connection.\n",
>>  			cmd->init_task_tag, na->dataout_timeout_retries);
>>  		goto failure;
>>
> 
> For this and the nop one I think we should print something so we know
> what session failed. For the nop one the sid can change on relogin if
> this was the only session, so you can't later match it from the logs and

I meant only connection in the session.

> configfs.
> 
> Maybe in the nop/dataout failure path add the I_T nexus values for the
> failed session in the failure message or maybe for both messages log the
> {sid, cid}  but then on login also log a message indicating a connection
> logged in and the I_T nexus to sid/cid mapping.
> 
>