Javen,
So i did try using iscsitadm and iscsiadm at
the initiator end just to see how the MPxIO works in
solaris.
At my target end ,have one working physical NIC
interface (there is another one but for some reason it
does not show up in prtconf as well ,so unable to
configure it) ..so i created 2 virtual interfaces to
this one interface having different ipaddresses say
10.10.10.226 and 10.10.10.227 .Using iscsitadm
,created 2 TPGT's (say 1 and 2 ) each assigned to one
of the above ips ,then i created a very simple target
of size 1g
On initiator side, through iscsiadm i was able to do a
sendtargets discovery on one of the ipaddresses above
and sure enough after the discovery when i did
#iscsiadm list target -v
it showed 2 connections one for each of the TPGT's
the 'format' command showed one disk scsi_vhci and
another regular device file ( this is not consistent
though ,one time it just showed just the one scsi_vhci
disk file )
in any case ,'mpathadm list lu' output indicates that
there are 2 paths
ALL i wanted to test was the 'failback' case ,so i ran
some I/O's on the scsi_vhci lun for a while ,then i
disabled ('down') one of the virtual interfaces on the
target m/c , NOW as per MPxIO i would expect that the
I/O's continue on normally as if nothing had happened
and fail over to the other virtual interface right ?
or is there a certain time period i should wait for
this to kick in ?
Either way i couldnt get failover mode to work ,not
sure if its something to do with the setup ,Pls let me
know
I have configured load_balance=none in my
scsi_vhci.conf on the initiator
Thanks
Som
--- Javen Wu <[EMAIL PROTECTED]> wrote:
> Hi Som,
>
> Since you don't know exact type of your iscsi
> target. So I think the
> best way to debug
> your iscsi initiator is to use Solaris iscsi target,
> please refer to
> iscsitadm(1M).
> The Solaris iscsi target supports ALUA, so it's easy
> to debug your iscsi
> initiator driver.
>
> Javen
>
> Javen Wu wrote:
>
> >Hi Som,
> >
> >The NV code of scsi_vhci is very different with S10
> version.
> >you are running S10. So please trace
> symmetric_device_probe() of
> >scsi_vhci module on S10.
> >
> >Javen
> >
> >Somnath kotur wrote:
> >
> >
> >
> >>Javen,
> >> I think my target is returning a valid GUID
> >>(Will investigate more on this and confirm )
> >>Oh and BTW i am using Solaris 10x86 off the CD -
> 08/07
> >>,does that make a difference?
> >>
> >>
> >>--- Javen Wu <[EMAIL PROTECTED]> wrote:
> >>
> >>
> >>
> >>
> >>
> >>>Hi Som,
> >>>
> >>>There are two conditions to enable mpxio.
> >>>
> >>>1. your target return a valid GUID.
> >>>2. your iscsi target can match at least one of
> below
> >>>failover types:
> >>> "misc/scsi_vhci/scsi_vhci_f_asym_sun",
> >>> "misc/scsi_vhci/scsi_vhci_f_asym_lsi",
> >>> "misc/scsi_vhci/scsi_vhci_f_asym_emc",
> >>> "misc/scsi_vhci/scsi_vhci_f_sym_emc",
> >>> "misc/scsi_vhci/scsi_vhci_f_sym_hds",
> >>> "misc/scsi_vhci/scsi_vhci_f_sym",
> >>> "misc/scsi_vhci/scsi_vhci_f_tpgs";
> >>>Oh, I assume you uses SXDE solaris rather than
> S10.
> >>>I paste the comments in scsi_vhci.conf, I think
> the
> >>>comments described
> >>>very clear:
> >>>
> >>>#
> >>># For a device that has a GUID, discovered on a
> pHCI
> >>>with mpxio enabled,
> >>>vHCI
> >>># access also depends on one of the scsi_vhci
> >>>failover modules accepting the
> >>># device. The default way this occurs is by a
> >>>failover module's "probe"
> >>># implementation (sfo_device_probe) indicating
> the
> >>>device is supported under
> >>># scsi_vhci. To override this default
> >>>probe-oriented configuration in
> >>># order to
> >>>#
> >>># 1) establish support for a device not
> currently
> >>>accepted under
> >>>scsi_vhci
> >>>#
> >>># or 2) override the module selected by "probe"
> >>>#
> >>># or 3) disable scsi_vhci support for a device
> >>>#
> >>># you can add a 'scsi-vhci-failover-override'
> tuple,
> >>>as documented in
> >>># scsi_get_device_type_string(9F). For each
> tuple,
> >>>the first part provides
> >>># basic device identity information (vid/pid) and
> >>>the second part selects
> >>># the failover module by "failover-module-name".
> If
> >>>you want to disable
> >>># scsi_vhci support for a device, use the special
> >>>failover-module-name
> >>>"NONE".
> >>># Currently, for each failover-module-name in
> >>>'scsi-vhci-failover-override'
> >>># (except "NONE") there needs to be a
> >>>#
> "misc/scsi_vhci/scsi_vhci_<failover-module-name>"
> >>>in 'ddi-forceload'
> >>>above.
> >>>#
> >>># " 111111"
> >>># "012345670123456789012345",
> >>>"failover-module-name" or "NONE"
> >>># "|-VID--||-----PID------|",
> >>># scsi-vhci-failover-override =
> >>># "STK FLEXLINE 400",
> >>>"f_asym_lsi",
> >>># "SUN T4", "f_tpgs",
> >>># "CME XIRTEMMYS", "NONE";
> >>>#
> >>>#END: FAILOVER_MODULE_BLOCK (DO NOT MOVE OR
> DELETE)
> >>>
> >>>If your mdi_pi_online() fails still, I think you
> >>>should trace the
> >>>*sfo_device_probe() for
> >>>scsi_vhci_f_sym module.
> >>>
> >>>I am not sure what kind of iscsi target you used?
> do
> >>>you used software
> >>>iscsi target?
> >>>If it is, I think it's symmetric. So please trace
>
> >>>symmetric_device_probe() of module
> >>>scsi_vhci_f_sym, see why you device_probe failed.
> >>>
> >>>Cheers
> >>>Javen
> >>>
> >>>
> >>>Somnath kotur wrote:
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>>Javen,
> >>>> Assuming that this is an iscsi target
> that
> >>>>
> >>>>
> >>>>
> >>>>
> >>>i
> >>>
> >>>
> >>>
> >>>
> >>>>have written/rather have full control over it,
> how
> >>>>
> >>>>
> >>>>
> >>>>
> >>>can
> >>>
> >>>
> >>>
> >>>
> >>>>i make sure that my iscsi initiator driver or
> even
> >>>>
> >>>>
> >>>>
> >>>>
> >>>the
> >>>
> >>>
> >>>
> >>>
> >>>>'iscsiadm' (Solaris's own iscsi initiator) is
> able
> >>>>
> >>>>
> >>>>
> >>>>
> >>>to
> >>>
> >>>
> >>>
> >>>
> >>>>use MPxIO on it?
> >>>>
> >>>>I have configured the target to have the same
> name
> >>>>with 2 portals, each with different target
> portal
>
=== message truncated ===
____________________________________________________________________________________
You rock. That's why Blockbuster's offering you one month of Blockbuster Total
Access, No Cost.
http://tc.deals.yahoo.com/tc/blockbuster/text5.com
_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss
