Hi All,

It seems there are still some problems, although possibly not with multipathing. Later on this evening when trying to do a simple zpool create the system panic'd. I thought that it may having something to do while using the multipathed device so i disabled multipathing and attempted creating a zpool using the first path and had pretty much the same panic. Then I attempted creating the zpool using the second path and had a different result. The zpool create command just hung. The box was still responsive so i could open another ssh session but any zfs/zpool/df just hung.

Fwiw, i was able to create a zpool without problems using the storedge 6140 multipathed device

I do have successful vmdumps for each of the panics and the panics are reproducible. I'll try to make these dumps available.

Im going to play with unmapping the volumes and remapping with different OS types. I see that the compellent storage center offers the following types:
Solaris 10 (what was currently used to generate the issues above)
Solaris 9
Solaris 8
Other Multipath
Other Singlepath
Other NoLUNGap
Compellent zNas 3.x (opensolaris based i think)

Then there are plenty of others (Linux, windows, HP UX, AIX, VMware, Xenserver etc..)

Here is the output from fmdump for the panic when attempting pool creation against the multipathed device

TIME UUID SUNW-MSG-ID Oct 18 2012 19:49:45.082601000 89ded12a-50e3-4c91-a7eb-a1567e954fd1 SUNOS-8000-KL

  TIME                 CLASS                                 ENA
Oct 18 19:49:45.0394 ireport.os.sunos.panic.dump_available 0x0000000000000000 Oct 18 19:49:33.2669 ireport.os.sunos.panic.dump_pending_on_device 0x0000000000000000

nvlist version: 0
        version = 0x0
        class = list.suspect
        uuid = 89ded12a-50e3-4c91-a7eb-a1567e954fd1
        code = SUNOS-8000-KL
        diag-time = 1350607785 51918
        de = fmd:///module/software-diagnosis
        fault-list-sz = 0x1
        fault-list = (array of embedded nvlists)
        (start fault-list[0])
        nvlist version: 0
                version = 0x0
                class = defect.sunos.kernel.panic
                certainty = 0x64
asru = sw:///:path=/var/crash/unknown/.89ded12a-50e3-4c91-a7eb-a1567e954fd1 resource = sw:///:path=/var/crash/unknown/.89ded12a-50e3-4c91-a7eb-a1567e954fd1
                savecore-succcess = 1
                dump-dir = /var/crash/unknown
                dump-files = vmdump.2
                os-instance-uuid = 89ded12a-50e3-4c91-a7eb-a1567e954fd1
panicstr = BAD TRAP: type=e (#pf Page fault) rp=ffffff002fd48880 addr=678 occurred in module "unix" due to a NULL pointer dereference panicstack = unix:die+ea () | unix:trap+1086 () | unix:cmntrap+e6 () | unix:mutex_enter+b () | zfs:zio_buf_alloc+25 () | zfs:zio_vdev_io_start+18a () | zfs:zio_execute+8b () | zfs:zio_nowait+21 () | zfs:vdev_probe+101 () | zfs:vdev_open+2e5 () | zfs:vdev_open_child+28 () | genunix:taskq_thread+2d9 () | unix:thread_start+8 () |
                crashtime = 1350607576
                panic-time = Thu Oct 18 19:46:16 2012 CDT
        (end fault-list[0])

        fault-status = 0x1
        severity = Major
        __ttl = 0x1
        __tod = 0x5080a3a9 0x4ec6428




I'll report back after I play with some other configurations

Thanks,
Majeed


Quoting Majeed <mab...@ilstu.edu>:

Thank you very much Joe, thats a win!

Multipathing against compellent now works. Format's output is condense and mpathadm list LU shows the paths.


% echo | format
Searching for disks...done

AVAILABLE DISK SELECTIONS:
0. c0t600A0B800033816200001391507FF856d0 <SUN-CSM200_R-0750 cyl 13052 alt 2 hd 255 sec 63>
          /scsi_vhci/disk@g600a0b800033816200001391507ff856
1. c0t6000D310004A06000000000000000014d0 <COMPELNT-CompellentVol-0505 cyl 65268 alt 2 hd 255 sec 63>
          /scsi_vhci/disk@g6000d310004a06000000000000000014

% mpathadm list lU
        /dev/rdsk/c0t6000D310004A06000000000000000014d0s2
                Total Path Count: 2
                Operational Path Count: 2
        /dev/rdsk/c0t600A0B800033816200001391507FF856d0s2
                Total Path Count: 4
                Operational Path Count: 4



Also, i did have device-type-scsi-options-list specified in the scsi_vhci.conf following the documentation as you mentioned. I commented that out and all is well! (it didn't seem to hurt having it there, but as you mentioned, definately wrong and not needed)

Thanks again,
Majeed

On 10/18/2012 04:12 PM, Joe Hartley wrote:
I just had this with another SAN vendor yesterday. The issue I ran into was that almost all the documentation I saw told me to modify /kernel/drv/scsi_vhci.conf by adding a device-type-scsi-options-list section, but that's wrong!

You need to modify that file, alright, but it needs a scsi-vhci-failover-override section, like this:

scsi-vhci-failover-override =
       "COMPELNTCompellent",        "f_sym";

Note that the second half of the identifier is the product name, but only up to the first space!

Good luck.

On Thu, Oct 18, 2012 at 4:49 PM, Majeed <mab...@ilstu.edu <mailto:mab...@ilstu.edu>> wrote:

   Hi All, As noted below, we have attempted to test the illumos
   based omniOS in our datacenter on a dell R610 connected through
   dual cisco fiber channel fabrics to a compellent SAN with dual
   controllers. We got pretty far with an out of the box setup until
   the last multipathing bit. I was hoping there might be some
   additional eyes on this list that may be able to provide some insight.

   Thanks in advance for taking the time to read through my following
   notes on this topic,
   Majeed Qulbain




   On 10/18/2012 03:06 PM, Majeed wrote:

       Thanks Theo, that makes sense. I will definitely do that.

       Majeed

       On 10/18/2012 02:58 PM, Theo Schlossnagle wrote:

           This is a more general Illumos problem, you will open your
           troubleshooting to a larger audience by interfacing with the
           illumos-discuss list.  I encourage you to post there! I'm
           subscribed
           there as well, but I haven't done multipathing against
           compellent and
           it appears no one else on the omnios-discuss list has either.
           Anything learned in the illumos-discuss group can be
           applied directly
           to OmniOS.

           On Thu, Oct 18, 2012 at 3:51 PM, Majeed<mab...@ilstu.edu
           <mailto:mab...@ilstu.edu>>  wrote:

               Hi Again, just wanted to update the list with my
               latest progress and
               thoughts on this topic. I think the issue lies in
               libmpscsi_vhci.so's lack
               of support for multipathing with compellent.

               When I do "mpathadm show mpath-support
               libmpscsi_vhci.so" on omniOS I see
               many venders and models listed but not compellent. We
               are going to spin up a
               solaris 11 machine to see what it looks like. I wonder
               if this was forked
               just before SUN/oracle added support for compellent (i
               could be way off base
               here).

               When I do the same command on solaris 10 u10, at the
               very bottom of the list
               i see.
                                Vendor:  COMPELNT
                                Product:  Compellent Vol
                                Revision:
                                Supported Load Balance Types:
                                        round-robin

               I have verified that multipathing is working properly
               on other arrays. I
               zoned up omniOS with all controllers on a 6140
               (CSM200) that sit in multiple
               fabrics, mapped a lun, and it looks wonderful. mpatham
               reports as expected,
               and all paths are condensed to a single block device
               in format's output.

               % mpathadm list LU
                    /dev/rdsk/c0t600A0B800033816200001391507FF856d0s2
                        Total Path Count: 4
                        Operational Path Count: 4



               It seems there are some options in
               /kernel/drv/scsi_vhci.conf to force
               scsi_vhci to accept 3rd party devices (hopefully
               compellent) for failover,
               but this is kinda getting over my head. It seems like
               if I had a proper
               ddi-forceload ''misc/scsi_vhci/scsi_vhci_f_<x>" file
               for compellent and I
               knew the proper VenderID and ProductID for the
               compellent devices, i might
               be able to force it. Im not sure if can use the
               information from the format
               inquiry to populate these values.

               format>  inquiry
               Vendor:   COMPELNT
               Product:  Compellent Vol
               Revision: 0505


               Any thoughts on this would be excellent.

               Thanks,
               Majeed Qulbain



               On 10/18/2012 09:30 AM, Majeed wrote:

                   I just noticed that we do have access to the
                   mpathadm command by
                   installing the
                   "pkg:/storage/mpathadm@0.5.11-0.151002" package. After
                   installing I don't see any paths with "mpathadm
                   list LU" which is what im
                   used to on solaris 10 (i know I should be
                   comparing to solaris 11, just
                   havn't played with it much). I wonder if there are
                   any other bits that I
                   need to install or other configuration tweaks to
                   get this going.

                   Thanks in advance for any pointers,
                   Majeed

                   Some other related commands that might assist
                   troubleshooting:

                   % uname -a
                   SunOS jhx011 5.11 omnios-eae537b i86pc i386 i86pc

                   % mpathadm list initiator-port
                   Initiator Port:  w5d4ae520937bbb00
                   Initiator Port:
iqn.1986-03.com.sun:01:b06a64f606ff.507f35b3,4000002a00ff
                   Initiator Port:  w5d4ae520937bbb00
                   Initiator Port:  2101001b32b40767
                   Initiator Port:  2100001b32940767

                   % luxadm -e port
/devices/pci@0,0/pci8086,3410@9/pci1077,143@0/fp@0,0:devctl
                   CONNECTED
/devices/pci@0,0/pci8086,3410@9/pci1077,143@0,1/fp@0,0:devctl
                   CONNECTED

                   % luxadm -v probe

                   Found Fibre Channel device(s):
                      Node WWN:5000d310004a0602  Device Type:Disk device
                        Logical Path:/dev/rdsk/c16t5000D310004A064Bd1s2
                      Node WWN:5000d310004a0602  Device Type:Disk device
                        Logical Path:/dev/rdsk/c17t5000D310004A064Cd1s2

                   % luxadm disp /dev/rdsk/c16t5000D310004A064Bd1s2
                   DEVICE PROPERTIES for disk:
                   /dev/rdsk/c16t5000D310004A064Bd1s2
                      Vendor:        COMPELNT
                      Product ID:        Compellent Vol
                      Revision:        0505
                      Serial Num:        00004a06-00000014
                      Unformatted capacity:    512000.000 MBytes
                      Read Cache:        Enabled
                        Minimum prefetch:    0x100
                        Maximum prefetch:    0xf0
                      Device Type:        Disk device
                      Path(s):

                      /dev/rdsk/c16t5000D310004A064Bd1s2

/devices/pci@0,0/pci8086,3410@9/pci1077,143@0/fp@0,0/disk@w5000d310004a064b,1:c,raw

                       Controller          /dev/cfg/c16
                        Device Address        5000d310004a064b,1
                        Host controller port WWN    2100001b32940767
                       Controller          /dev/cfg/c17
                        Device Address        5000d310004a064c,1
                        Host controller port WWN    2101001b32b40767


                   On 10/17/2012 06:25 PM, Majeed wrote:

                       Hi All, We are doing some testing with omniOS
                       in our datacenter and would
                       love connect it up to our compellent san via
                       fc connectivity. I slapped
                       omniOS onto a dell r610 that has a qlogic dual
                       port hba (each port is pathed
                       to separate fabrics), performed zoning in each
                       fabric, and assigned some
                       luns. Now im not sure if compellent supports
                       Solaris 11 (i can contact them
                       and see about that), but it definately support
                       solaris 10 which is what we
                       use on other production systems.

                       I'v actually made it pretty far through the
                       process which is really
                       exciting, except for the last multipathing
                       bit. After mapping a volume to
                       the host I can't seem to condense all the
                       devices down to a single
                       mulipathed device. It seems multipathing is
                       enabled for the fp driver. I
                       noticed there is not a mpathadm command.

                       Anyone have any ideas? I'll post some output
                       from the commands i think of
                       below. Let me know if you would like to see
                       any other information.

                       Thanks much,
                       Majeed


                       On solaris 10 the main thing we have to do is
                       add the following to
                       /kernel/drv/scsi_vhci.conf (reconfigure reboot)
                       device-type-scsi-options-list =
                       "COMPELNTCompellent Vol", "symmetric-option";
                       symmetric-option = 0x1000000;


                       Here is the format output:
                       AVAILABLE DISK SELECTIONS:
                               0.
                       c13t5000CCA0213ED1E1d0<HITACHI-HUC106030CSS600-A360
                       cyl 36470
                       alt 2 hd 255 sec 63>

/pci@0,0/pci8086,2940@1c/pci1028,1f1e@0/iport@40/disk@w5000cca0213ed1e1,0

                               1.
                       c14t5000CCA0213EFC7Dd0<HITACHI-HUC106030CSS600-A360
                       cyl 36469
                       alt 2 hd 255 sec 63>

/pci@0,0/pci8086,2940@1c/pci1028,1f1e@0/iport@80/disk@w5000cca0213efc7d,0

                               2.
                       c16t5000D310004A064Bd1<COMPELNT-CompellentVol-0505
                       cyl 65268
                       alt 2 hd 255 sec 63>

/pci@0,0/pci8086,3410@9/pci1077,143@0/fp@0,0/disk@w5000d310004a064b,1
                               3.
                       c17t5000D310004A064Cd1<COMPELNT-CompellentVol-0505
                       cyl 65268
                       alt 2 hd 255 sec 63>

/pci@0,0/pci8086,3410@9/pci1077,143@0,1/fp@0,0/disk@w5000d310004a064c,1

                       Specify disk (enter its number): Specify disk
                       (enter its number):



                       % fcinfo hba-port
                       HBA Port WWN: 2100001b32940767
                            Port Mode: Initiator
                            Port ID: 91c00
                            OS Device Name: /dev/cfg/c16
                            Manufacturer: QLogic Corp.
                            Model: 375-3356-02
                            Firmware Version: 05.03.01
                            FCode/BIOS Version:  BIOS: 2.02; fcode:
                       2.01; EFI: 2.00;
                            Serial Number: 0402H00-0941774287
                            Driver Name: qlc
                            Driver Version: 20100408-3.01
                            Type: N-port
                            State: online
                            Supported Speeds: 1Gb 2Gb 4Gb
                            Current Speed: 4Gb
                            Node WWN: 2000001b32940767
                            Max NPIV Ports: 127
                            NPIV port list:
                       HBA Port WWN: 2101001b32b40767
                            Port Mode: Initiator
                            Port ID: a1900
                            OS Device Name: /dev/cfg/c17
                            Manufacturer: QLogic Corp.
                            Model: 375-3356-02
                            Firmware Version: 05.03.01
                            FCode/BIOS Version:  BIOS: 2.02; fcode:
                       2.01; EFI: 2.00;
                            Serial Number: 0402H00-0941774287
                            Driver Name: qlc
                            Driver Version: 20100408-3.01
                            Type: N-port
                            State: online
                            Supported Speeds: 1Gb 2Gb 4Gb
                            Current Speed: 4Gb
                            Node WWN: 2001001b32b40767
                            Max NPIV Ports: 127
                            NPIV port list:


                       % cfgadm -al -o show_SCSI_LUN
Ap_Id Type Receptacle Occupant
                       Condition
c16 fc-fabric connected configured
                       unknown
c16::5000d310004a0649 unknown connected unconfigured
                       unknown
c16::5000d310004a064b,1 disk connected configured
                       unknown
c17 fc-fabric connected configured
                       unknown
c17::5000d310004a064a unknown connected unconfigured
                       unknown
c17::5000d310004a064c,1 disk connected configured
                       unknown


                       _______________________________________________

   --     Majeed Abu-Qulbain
   Lead Unix Admin - CISS
   309-438-3455 <tel:309-438-3455>
   Illinois State University



   -------------------------------------------
   illumos-discuss
   Archives: https://www.listbox.com/member/archive/182180/=now
   RSS Feed:
   https://www.listbox.com/member/archive/rss/182180/22222335-76d00b28
   Modify Your Subscription:
   https://www.listbox.com/member/?&;
   <https://www.listbox.com/member/?&;>
   Powered by Listbox: http://www.listbox.com




--
Joe Hartley, Software Engineer
GreenBytes, Inc.
401.315.5580

*illumos-discuss* | Archives <https://www.listbox.com/member/archive/182180/=now> <https://www.listbox.com/member/archive/rss/182180/23508103-b88c366c> | Modify <https://www.listbox.com/member/?&;> Your Subscription [Powered by Listbox] <http://www.listbox.com>



--
Majeed Abu-Qulbain
Lead Unix Admin - CISS
309-438-3455
Illinois State University




-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/23508103-b88c366c
Modify Your Subscription: https://www.listbox.com/member/?&;
Powered by Listbox: http://www.listbox.com




----------------------------------------------------------------
This message was sent using Illinois State University RedbirdMail




-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com

Reply via email to