[dm-devel] [PATCH] multipath: retry the DID_SOFT_ERROR for rdac checker commands

[Christophe Varoqui]

On mar., 2012-07-10 at 19:12 +0000, Moger, Babu wrote:
> Sometimes we have seen immediate path failures for DID_SOFT_ERROR status. Just add a retry as other statuses.
> It will basically add 5 retries.
> 
Applied.
Thanks.

> Here are the messages.
> Jun  4 17:46:42 ictc-billy kernel: mpt2sas0: log_info(0x31120100): originator(PL), code(0x12), sub_code(0x0100)
> Jun  4 17:46:42 ictc-billy kernel: sd 1:0:0:47: [sdn] Unhandled error code
> Jun  4 17:46:42 ictc-billy kernel: sd 1:0:0:47: [sdn] Result: hostbyte=DID_SOFT_ERROR driverbyte=DRIVER_OK
> Jun  4 17:46:42 ictc-billy kernel: sd 1:0:0:47: [sdn] CDB: Write(10): 2a 00 00 06 e4 a0 00 00 20 00
> Jun  4 17:46:42 ictc-billy kernel: device-mapper: multipath: Failing path 8:208.
> Jun  4 17:46:42 ictc-billy multipathd: 8:208: mark as failed
> Jun  4 17:46:42 ictc-billy multipathd: mpathat: remaining active paths: 1
> 
> 
> Signed-off-by: Babu Moger <babu.moger at netapp.com>
> ---
> --- multipath-tools/libmultipath/checkers/rdac.c.orig	2012-07-10 13:35:34.000000000 -0500
> +++ multipath-tools/libmultipath/checkers/rdac.c	2012-07-10 13:36:07.000000000 -0500
> @@ -166,6 +166,7 @@ retry:
>  		switch (io_hdr.host_status) {
>  		case DID_BUS_BUSY:
>  		case DID_ERROR:
> +		case DID_SOFT_ERROR:
>  		case DID_TRANSPORT_DISRUPTED:
>  			/* Transport error, retry */
>  			if (--retry_rdac)
> 
>