Secondary fiber card not staying enabled

[superla...@gmail.com]

I installed ESOS and I had everything configured and looking good, I sync'd the config and to perform a final reboot test, i rebooted the unit only to watch the unit boot up with my secondary fiber card being disabled and the config info not being loaded. I have verified that I have the correct copy of scst.conf on both the /etc and on the USB flash. How do I keep ESOS from disabling my secondary fiber card? 

I even went as far as removing the shutdown sync to prevent any changes to the config file and yet the unit is still not rebooting properly. I cannot use this in any kind of production environment for my ESXi infrastructure if I can't reliably get my storage to map correctly across the FC Cards.

Any suggestions?

Thank you

[Marc Smith]

Hi,

This is the first time I've heard of this type of behavior reported...
first, I assume you are using version 0.1.9 (from 0.1.x)? Can you send
the output of the "dmesg" command, the /var/log/boot file, and the
/etc/scst.conf file?


--Marc

> --
> You received this message because you are subscribed to the Google Groups
> "esos-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to esos-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[superla...@gmail.com]

First of all I would like to thank you for your assistance.

Sure! Attached you should find a copy of the scst.conf file as well as the dmesg output.

Everything seems to stay operational until i reboot ESOS and that is when my secondary card comes up as being disabled although the conf file shows it as being enabled. I even went as far as making a master copy and overwriting the scst.conf at boot but it doesn't seem to make any difference.

The physical server in question is a Dell R710 with six 1TB SATA drives configured as three sets of mirrors which I labeled as being FC_VOL1, FC_VOL2, and FC_VOL3. FC_VOL1 is mapped to a Windows VM (Px_BACKUP) and FC_VOL2 and FC_VOL3 are mapped to Px_VMWH1, Px_VMWH2, and Px_VMWH3, which are VMware ESXi hosts. The FC switch is a Cisco MDS 9124.

Thank you

Sincerely,

John Strader

[superla...@gmail.com]

The version of ESOS implemented is 0.1.9

thank you

Sincerely,

John Strader

On Tuesday, November 15, 2016 at 12:15:11 PM UTC-8, superla...@gmail.com wrote:

[Marc Smith]

So, I looked at your logs, and take a look at these lines from the kernel log:

--snip--
[ 78.838369] qla2xxx [0000:00:00.0]-0005: QLogic Fibre Channel HBA
Driver: 8.04.00.05.2.2-SCST.03-k-p.
[ 78.838529] qla2xxx [0000:04:00.0]-001d: Found an ISP2432 irq 34
iobase 0xffffc9001127c000.
[ 78.839073] qla2xxx [0000:04:00.0]-0034:7: MSI-X: Unsupported ISP
2432 SSVID/SSDID (0x103C,0x7041).
[ 79.149758] scsi7 : qla2xxx
[ 79.561779] qla2xxx [0000:04:00.0]-505f:7: Link is operational (4 Gbps).
[ 79.943983] bnx2 0000:01:00.1 eth1: NIC Copper Link is Up, 1000
Mbps full duplex
[ 79.943987] , receive & transmit flow control ON
[ 79.944072] IPv6: ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[ 79.988993] bnx2 0000:02:00.0 eth2: NIC Copper Link is Up, 1000
Mbps full duplex
[ 79.988998] , receive & transmit flow control ON
[ 79.989088] IPv6: ADDRCONF(NETDEV_CHANGE): eth2: link becomes ready
[ 80.119318] qla2xxx [0000:04:00.0]-00fb:7: QLogic HPAE312A -
PCI-Express Dual Port 4Gb Fibre Channel HBA.
[ 80.119333] qla2xxx [0000:04:00.0]-00fc:7: ISP2432: PCIe (2.5GT/s
x4) @ 0000:04:00.0 hdma+ host#=7 fw=7.03.00 (9496).
[ 80.119453] qla2xxx [0000:04:00.1]-001d: Found an ISP2432 irq 44
iobase 0xffffc90011464000.
[ 80.120044] qla2xxx [0000:04:00.1]-0034:8: MSI-X: Unsupported ISP
2432 SSVID/SSDID (0x103C,0x7041).
[ 80.160089] bnx2 0000:02:00.1 eth3: NIC Copper Link is Up, 1000
Mbps full duplex
[ 80.160093] , receive & transmit flow control ON
[ 80.160172] IPv6: ADDRCONF(NETDEV_CHANGE): eth3: link becomes ready
[ 80.276622] bnx2 0000:01:00.0 eth0: NIC Copper Link is Up, 1000
Mbps full duplex
[ 80.276628] , receive & transmit flow control ON
[ 80.276712] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 80.430801] scsi8 : qla2xxx
[ 85.440296] qla2xxx [0000:04:00.1]-00fb:8: QLogic HPAE312A -
PCI-Express Dual Port 4Gb Fibre Channel HBA.
[ 85.440311] qla2xxx [0000:04:00.1]-00fc:8: ISP2432: PCIe (2.5GT/s
x4) @ 0000:04:00.1 hdma+ host#=8 fw=7.03.00 (9496).
[ 85.440430] qla2xxx [0000:05:00.0]-001d: Found an ISP2432 irq 35
iobase 0xffffc9001160a000.
[ 85.441096] qla2xxx [0000:05:00.0]-0034:9: MSI-X: Unsupported ISP
2432 SSVID/SSDID (0x103C,0x7041).
[ 85.751765] scsi9 : qla2xxx
[ 86.182741] qla2xxx [0000:05:00.0]-505f:9: Link is operational (4 Gbps).
[ 86.755262] qla2xxx [0000:05:00.0]-00fb:9: QLogic HPAE312A -
PCI-Express Dual Port 4Gb Fibre Channel HBA.
[ 86.755276] qla2xxx [0000:05:00.0]-00fc:9: ISP2432: PCIe (2.5GT/s
x4) @ 0000:05:00.0 hdma+ host#=9 fw=7.03.00 (9496).
[ 86.755394] qla2xxx [0000:05:00.1]-001d: Found an ISP2432 irq 46
iobase 0xffffc900117b0000.
[ 86.756051] qla2xxx [0000:05:00.1]-0034:10: MSI-X: Unsupported ISP
2432 SSVID/SSDID (0x103C,0x7041).
[ 87.066615] scsi10 : qla2xxx
[ 92.076239] qla2xxx [0000:05:00.1]-00fb:10: QLogic HPAE312A -
PCI-Express Dual Port 4Gb Fibre Channel HBA.
[ 92.076253] qla2xxx [0000:05:00.1]-00fc:10: ISP2432: PCIe (2.5GT/s
x4) @ 0000:05:00.1 hdma+ host#=10 fw=7.03.00 (9496).
--snip--

And then below that, I see lines like this:

--snip--
[ 102.492018] qla2xxx [0000:04:00.1]-8038:8: Cable is unplugged...
[ 109.127063] qla2xxx [0000:05:00.1]-8038:10: Cable is unplugged...
[ 110.129692] qla2xxx [0000:05:00.1]-00af:10: Performing ISP error
recovery - ha=ffff880202911000.
[ 110.169153] qla2xxx [0000:05:00.1]-1073:10: Disabling FC tape support
[ 119.051852] qla2x00t(9): local session for loop_id 130 deleted
[ 119.055043] qla2x00t(9): local session for loop_id 132 deleted
[ 130.906229] qla2xxx [0000:05:00.1]-8038:10: Cable is unplugged...
[ 131.332201] qla2xxx [0000:05:00.0]-505f:9: Link is operational (4 Gbps).
--snip--

So, I'd say something is up... is this one of those 4-port QLogic FC
HBA's (4 ports on one card)? Or maybe this an HP-branded HBA? That
shouldn't really matter though... first thing I try if you have it is
a different HBA.

This line is interesting too:
[ 86.756051] qla2xxx [0000:05:00.1]-0034:10: MSI-X: Unsupported ISP
2432 SSVID/SSDID (0x103C,0x7041).

Actually, I just did a quick search and there is even a post on the
scst-devel list about it:
https://sourceforge.net/p/scst/mailman/message/34541780/

Sounds like a couple options then: Use a different QLogic HBA, or
possibly build ESOS using the QLogic-provided target driver (via the
--enable-qla option). I would lean towards using a different HBA,
preferably an OEM, or at least a normal single-port or dual-port card.
In my experience (and others) the QLogic-provided driver is not as
stable as the SCST-provided driver (default).

Let us know how it goes.


--Marc

[superla...@gmail.com]

Marc,

I decided to redo the USB stick image and start from scratch to see how the HBA adapters behaved from first boot without anything configured. I went into the bios on each adapter and made sure each adapter was reset to manufacturer defaults and then I enabled LIP support, saved and rebooted.

At first boot of ESOS all HBA adapters were disabled as expected.

The first thing i did was set the date/time and sync'd that configuration and rebooted.

Next I enabled all of the targets for the FC cards/ports

and finally I configured the IP address of the SAN.

Sync'd the configuration and rebooted.

The unit comes up at this point with all four targets as being enabled.

I will attempt to continue configuring the storage and mapping them to the FC names and see what happens.

I'll let you know the results.

Thank you

Sincerely,

John Strader

Superlative IT

On Tuesday, November 15, 2016 at 12:15:11 PM UTC-8, superla...@gmail.com wrote: