[sles-beta] SLES11-SP4 multipath problems, SAN path does not get reinstated after path failure

[Hannes Reinecke]

On 03/10/2015 11:34 AM, urs.frey at post.ch wrote:
> Hi all
> 
> I have installed SLES11-SP4 x86_64 on a HP Proliant BL460cgen8 attached to SAN EMC, FCoE.
> 
> When failing a SAN path, I can observe, that the SAN path does not get reinstated after SAN link up again.
> Multipath tools do loose all LUNs
> 
> h05cni:~ # multipath -ll
> 360000970000295700687533031374132 dm-2
> size=34G features='0' hwhandler='0' wp=rw
> 360000970000295700687533031374131 dm-1
> size=34G features='0' hwhandler='0' wp=rw
> 360000970000295700687533031374130 dm-0
> size=34G features='0' hwhandler='0' wp=rw
> 360000970000295700687533031443339 dm-3
> size=34G features='0' hwhandler='0' wp=rw
> h05cni:~ #
> 
> Mar 10 14:14:11 h05cni kernel: [79258.113860] lpfc 0000:04:00.3: 1:(0):0203 Devloss timeout on WWPN 50:00:09:75:00:0a:bd:a1 NPort x5c0080 Data: x0 x8 x3
> Mar 10 14:14:11 h05cni kernel: [79258.113900] lpfc 0000:04:00.3: 1:(0):0203 Devloss timeout on WWPN 50:00:09:75:00:0a:bd:18 NPort x5c0060 Data: x0 x8 x2
> Mar 10 14:16:50 h05cni kernel: [79417.676167] lpfc 0000:04:00.2: 0:1305 Link Down Event x3 received Data: x3 x20 x800110 x0 x0
> Mar 10 14:17:11 h05cni kernel: [79438.138987] lpfc 0000:04:00.2: 0:1303 Link Up Event x8 received Data: x8 x0 x40 x0 x0 x0 0
> Mar 10 14:17:11 h05cni kernel: [79438.139749] lpfc 0000:04:00.2: 0:2546 New FCF event, evt_tag:x9, index:x0
> Mar 10 14:17:21 h05cni kernel: [79447.751373] lpfc 0000:04:00.2: 0:(0):0203 Devloss timeout on WWPN 50:00:09:75:00:0a:bd:90 NPort xc00060 Data: x0 x8 x2
> Mar 10 14:17:21 h05cni kernel: [79447.751413] lpfc 0000:04:00.2: 0:(0):0203 Devloss timeout on WWPN 50:00:09:75:00:0a:bd:29 NPort xc00080 Data: x0 x8 x3
> 
That looks like an issue with the driver. If the devices disappear after
a LIP the driver isn't able to correctly attach the devices again.
Please open a bugzilla to track this properly.

Cheers,

Hannes