VIMA esxcfg-mpath -a bug

[mcowger]

So there is a bug in the VIMA 1.0. When you run esxcfg-mpath -a on a host with a service console, it out puts the WWPNs on the FC cards (good). When you run it on the VIMA against a host, you get the WWNN (bad).

Here is a diff to fix it:

sub list_hbas {

my ($ss, $lunName) = @_;

my $hbas = $ss->storageDeviceInfo->hostBusAdapter;

foreach my $hba (@$hbas) {

if (defined($hba)) {

next if $hba->isa("HostBlockHba");

next if $hba->isa("HostParallelScsiHba");

my $pciString = get_pci_string($hba);

printf("%s %s%s\n",

$hba->device,

-$hba->isa("HostFibreChannelHba") ? $hba->nodeWorldWideName . " " : "",

+$hba->isa("HostFibreChannelHba") ? $hba->portWorldWideName . " " : "",

$pciString);

}

}

}

Is this the right place for asuch a bug report?

--Matt

VCP, vExpert, Unix Geek

[mcowger]

It also outputs as a long, not a hex as in the one form a service console. long isn't very useful.....

--Matt

VCP, vExpert, Unix Geek

[lamw]

Hey Matt,

I think this is a good place for the bug, but if Harvey (PM for vMA/VIMA) does not get back you could also file an SR with Vmware)

Not sure if this was caught before or internally, but in the next release of vMA, the --a option is not longer available with esxcfg-mpath and you'll see some different format output which makes things easier to read. I know some other bugs there were found may not get fixed in 1.0 release and its been recommended that when vMA 4.0 is out, to just upgrade.

Here's a sample out from vMA with esxcfg-mpath -l:

[vi-admin@rafaeli ~]$ esxcfg-mpath --server everest.primp-industries.com -l
block.cciss/c0d0:0-block.0:0-mpx.vmhba0:C0:T0:L0
   Runtime Name: vmhba0:C0:T0:L0
   Device: mpx.vmhba0:C0:T0:L0
   Device Display Name: Local VMware Disk (mpx.vmhba0:C0:T0:L0)
   Adapter: vmhba0 Channel: 0 Target: 0 LUN: 0
   Adapter Identifier: block.cciss/c0d0:0
   Target Identifier: block.0:0
   Plugin: NMP
   State: active
   Transport: block

fc.20000000c97baa38:10000000c97baa38-fc.50014380013c90e0:50014380013c90ed-naa.50014380013c90e0
   Runtime Name: vmhba1:C0:T1:L0
   Device: naa.50014380013c90e0
   Device Display Name: HP Fibre Channel RAID Ctlr (naa.50014380013c90e0)
   Adapter: vmhba1 Channel: 0 Target: 1 LUN: 0
   Adapter Identifier: fc.20000000c97baa38:10000000c97baa38
   Target Identifier: fc.50014380013c90e0:50014380013c90ed
   Plugin: NMP
   State: active
   Transport: fc
   Adapter Transport Details: WWPN: 10:00:00:00:c9:7b:aa:38 WWNN: 20:00:00:00:c9:7b:aa:38
   Target Transport Details: WWPN: 50:01:43:80:01:3c:90:ed WWNN: 50:01:43:80:01:3c:90:e0

=========================================================================

William Lam

VMware vExpert 2009

VMware ESX/ESXi scripts and resources at:

If you find this information useful, please award points for "correct" or "helpful".

[mcowger]

Yeah, but if I don’t have any luns yet..., -l doesn't do much for me.

--M

[lamw]

True, looks like -a option has moved to what used to be the esxcfg-vmhbadevs command from what I can tell:

[root@everest ~]# esxcfg-scsidevs
esxcfg-scsidevs <options>
Print the mappings between vmhba names and /dev names
-l|--list                  List all Logical Devices known on this
                           system with device information.
-c|--compact-list           List all Logical Devices each on a
                           single line, with limited information.
-u|--uids                  List all Device Unique Identifiers with their primary name.
-d|--device                Used to filter the --list, --compact-list
                           and uids commands to limit output to a
                           specific device.
-m|--vmfs                  Print mappings for VMFS volumes to their
                           Service Console partitions and vmhba names.
-f|--vfat                  Print mappings for VFAT volumes to their
                           Service Console partitions and vmhba names.
-a|--hbas                  Print HBA devices with identifying information
-A|--hba-device-list       Print a mapping between HBAs and the devices it
                           provides paths to.
-o|--offline-cos-dev       Offline the COS device corresponding to this vmkernel
                           device.
-n|--online-cos-dev        Bring online the COS device corresponding to this vmkernel
                           device.
-h|--help                  Show this message.

[root@everest ~]# esxcfg-scsidevs -a
vmhba0  cciss             link-n/a  block.cciss/c0d0:0                      (11:8.0) Hewlett-Packard Company Smart                       Array E200i
vmhba1  lpfc820           link-up   fc.20000000c97baa38:10000000c97baa38    (16:0.0) Emulex Corporation LPe11000 4G                      b Fibre Channel Host Adapter
vmhba2  lpfc820           link-up   fc.20000000c97baa39:10000000c97baa39    (16:0.1) Emulex Corporation LPe11000 4G                      b Fibre Channel Host Adapter

=========================================================================

William Lam

VMware vExpert 2009

VMware ESX/ESXi scripts and resources at:

If you find this information useful, please award points for "correct" or "helpful".

[lamw]

FYI - this issue looks like it was fixed in the upcoming vMA release.

There will be a backwards compatible esxcfg-mpath for ESX 3.5 called esxcfg-mpath35 and the code has been updated:

sub list_hbas {
   my ($ss, $lunName) = @_;
   my $hbas = $ss->storageDeviceInfo->hostBusAdapter;
   foreach my $hba (@$hbas) {
      if (defined($hba)) {
         next if $hba->isa("HostBlockHba");
         next if $hba->isa("HostParallelScsiHba");
         my $pciString = get_pci_string($hba);
         printf("%s %s%s\n",
                $hba->device,
                $hba->isa("HostFibreChannelHba") ? dec_to_hex($hba->portWorldWideName) . " " : "",
                $pciString);
      }
   }
}

I believe there will not be any fixes/updates going to VIMA 1.0 and VMware will just recommend everyone to upgrade to vMA 4.0 when it's out

=========================================================================

William Lam

VMware vExpert 2009

VMware ESX/ESXi scripts and resources at:

If you find this information useful, please award points for "correct" or "helpful".

[halcabes]

lamw wrote:

Not sure if this was caught before or internally, but in the next release of vMA, the --a option is not longer available with esxcfg-mpath and you'll see some different format output which makes things easier to read. I know some other bugs there were found may not get fixed in 1.0 release and its been recommended that when vMA 4.0 is out, to just upgrade.

That sounds right.

Please try it with vSphere CLI 4.0 which will be included in vMA 4.0 (available later this month). If you encounter problems with that version please post them here. Thanks.

--Harvey

Harvey Alcabes

Product Manager

VMware, Inc.