|
Home > Archive > Data Storage > November 2004 > SAN or FC Switch not passing device "name" to HBA?
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
SAN or FC Switch not passing device "name" to HBA?
|
|
|
| Hi,
I realize that this is a kind of esoteric question, but I hope that
someone here can help.
We have some systems with dual-channel QLogic QLA2312 FC HBAs. These
are connected to an EMC 32-port FC Switch, which is in turn connected to
an EMC Clariion CX700 SAN.
Once we get everything configured correctly, the system can communicate
with the SAN/SAN LUN, etc. We are booting these systems from SAN LUNs,
so to do this configuration, part of what we do is to go into the BIOS
for the system to configure the boot order so that we boot from the "SAN
device" first.
When we do this, in some cases, the BIOS is showing a null entry "[]"
for the SAN "boot device", whereas in other cases, the entry in the boot
choice for the SAN "boot device" we get is showing as something like
"[HA...."]. In either case (whether the boot device is showing as "[]"
or "[HA....]", selecting the SAN boot device seems to work ok (i.e., the
system boots from the SAN LUN.
Now, we're been trying to figure out why. At first we thought that it
might be something to do with the system (e.g., motherboard or system
BIOS) or the HBA BIOS, but in all cases, all the systems are similar.
The system vendor is conjecturing that either the FC switch or the SAN
is suppose to send a string/identifier to the HBA, and that's why in
some cases, we get the "[]", whereas in other cases, we get the
"[HA....]".
We've been searching for awhile but as far as we can tell all the LUNs
associated with each of the systems are generally configured identically
except for possibly their size.
Does anyone know if this identifier/string might be something that the
FC switch (I think it's like a "DSM-32", and have been told that it's
actually a McData) might be configured to send or not to send, and if
so, any ideas of how to do this (we'd like to get the "[HA....]" in all
cases).
Thanks,
Jim
| |
|
| Hi,
In case anyone ever runs across this problem/question, I was able to
determine what can trigger this problem. During some testing today, it
appears that if only one of the HBA ports is enabled, the HBA appears as
"[]" in the BIOS, but if both HBA ports are enabled, the HBA appears as
"[HA...]". I haven't been able to determine if this is a bug or not in
the QLogic driver yet  ...
Jim
ohaya wrote:
>
> Hi,
>
> I realize that this is a kind of esoteric question, but I hope that
> someone here can help.
>
> We have some systems with dual-channel QLogic QLA2312 FC HBAs. These
> are connected to an EMC 32-port FC Switch, which is in turn connected to
> an EMC Clariion CX700 SAN.
>
> Once we get everything configured correctly, the system can communicate
> with the SAN/SAN LUN, etc. We are booting these systems from SAN LUNs,
> so to do this configuration, part of what we do is to go into the BIOS
> for the system to configure the boot order so that we boot from the "SAN
> device" first.
>
> When we do this, in some cases, the BIOS is showing a null entry "[]"
> for the SAN "boot device", whereas in other cases, the entry in the boot
> choice for the SAN "boot device" we get is showing as something like
> "[HA...."]. In either case (whether the boot device is showing as "[]"
> or "[HA....]", selecting the SAN boot device seems to work ok (i.e., the
> system boots from the SAN LUN.
>
> Now, we're been trying to figure out why. At first we thought that it
> might be something to do with the system (e.g., motherboard or system
> BIOS) or the HBA BIOS, but in all cases, all the systems are similar.
> The system vendor is conjecturing that either the FC switch or the SAN
> is suppose to send a string/identifier to the HBA, and that's why in
> some cases, we get the "[]", whereas in other cases, we get the
> "[HA....]".
>
> We've been searching for awhile but as far as we can tell all the LUNs
> associated with each of the systems are generally configured identically
> except for possibly their size.
>
> Does anyone know if this identifier/string might be something that the
> FC switch (I think it's like a "DSM-32", and have been told that it's
> actually a McData) might be configured to send or not to send, and if
> so, any ideas of how to do this (we'd like to get the "[HA....]" in all
> cases).
>
> Thanks,
> Jim
| |
|
| That's interesting. I hadn't seen anything like that. The "HA..." may
stand for high availability; maybe the qlogic driver is realizing that
there are two devices with the same LUN on each path. Just a guess...
I don't know what the qlogic card would do with that information,
though.
| |
|
|
AWS wrote:
>
> That's interesting. I hadn't seen anything like that. The "HA..." may
> stand for high availability; maybe the qlogic driver is realizing that
> there are two devices with the same LUN on each path. Just a guess...
> I don't know what the qlogic card would do with that information,
> though.
Hi,
I wrote to QLogic via their website, but haven't heard anything back
from them (yet).
Jim
| |
|
|
ohaya wrote:
>
> AWS wrote:
>
> Hi,
>
> I wrote to QLogic via their website, but haven't heard anything back
> from them (yet).
>
> Jim
Hi,
Just got an email from QLogic, and unfortunately, they're saying that
any support needs to come from the system vendor, whose not been able to
provide any info on this...
Jim
|
|
|
|
|