I didn't jump at the first posting because you mentioned it was emulex. But
seeing it's actually a QLogic, in my opinion you probably just need to run
devfsadm -Cv and then it should show up.
--
This message posted from opensolaris.org
___
The QLogic iSCSI HBA drivers do not support MPxIO. You would need the Symantec
software to handle host side failover. Now there is the basic version which may
meet your needs:
http://www.symantec.com/business/storage-foundation-basic
When looking at the download page it lists the following
On windows, QLogic used to have a filter driver called qldirect. That was
discontinued around 2004 in favor of Microsoft's MPIO which requires storage
vendors to provide a DSM.
--
This message posted from opensolaris.org
___
storage-discuss mailing
Back to this original issue, Have you followed all of the storage manufacturers
advice when changing the data rates and connection mode?
I have to ask as another customer I dealt with was having a data rate issue and
needed to change some dip switches on the storage and then reboot it to get
The client side should be plug and play.
Using the cfgadm -al command you should see a listing with the FC controller
port and when attached/logged into the switch the type will say fc-fabric.
Once it's logged into the switch, then the hba wwpn and storage wwpn need to be
together in a zone,
Here is the bug for the cause of the issue:
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6889922
Here are the bugs to work around it:
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6893529
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6857553
Here are the
How do you test different firmware on the QLE2462 with the qlt driver?
--
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
What tools will you use to evaluate the IOPS in this configuration?
--
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
I'm curious, what kind of switch are you using in this configuration?
--
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
Being I had library troubles with the pre-compiled versions of diskomizer, I
went though the following steps to get it going:
Download and Install the latest Sun compilers. How to do this is not included
here.
Download the latest Solaris Test Collection Source Code from the next URL to
the
With VTS installed run the following on the command line:
# /usr/sunvts/bin/invokeVtsAgent
# /usr/sunvts/bin/vts_cmd reset
# /usr/sunvts/bin/vts_cmd set_mode [Component Stress]
# /usr/sunvts/bin/vts_cmd disable_test all
# /usr/sunvts/bin/vts_cmd enable_test HBA
# /usr/sunvts/bin/vts_cmd
I've run into this on my amd64 host when trying to run diskomizer:
root:= bin/diskomizer -f test_options
Copyright 2009 Sun Microsystems, Inc. All Rights Reserved
Use is subject to license terms.
Version 8.0
/opt/SUNWstc-diskomizer/bin/amd64-xO4/diskomizer -f test_options
Run this command:
echo *stmf_trace_buf/s | mdb -k /tmp/filename.txt
Then attach the output file.
Are the symptoms similar to this:
http://bugs.opensolaris.org/bugdatabase/view_bug.do;jsessionid=aa858e69925c402fe4c0b8eb7bf7?bug_id=6804502
If it is that bug the only choice is to reboot the
Could you post the full stmf trace buffer here?
--
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
I doubt it's the firmware level running on the cards. But it could be the FC
card driver or the upper layer drivers or just the configuration of the host.
Does the storage show up using cfgadm? If it does you may need to use cfgadm
(man cfgadm_fp) to configure it.
--
This message posted from
The HBA is a 2GB HBA. Only the 4GB and 8GB HBAs from QLogic are supported with
the qlt driver.
--
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
Files in Celeste are versioned, and have audit trails.
This is SO cool! I miss the old VMS file versioning.
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
Well you should have a /etc/driver_classes entry with fibre-channel for qlc.
Then in the driver_aliases there should be a qlc entry with 1077,2200. That
should be about what's needed for the qlc driver to attach to the card.
This message posted from opensolaris.org
If you want to investigate why it is not working, edit the
/kernel/drv/qlc.conf, enable the extended logging and reboot. Then look in the
/var/adm/messages file to see what happens at driver attach time.
This message posted from opensolaris.org
___
You might be able to gleen some info using the scsi.d dtrace script from here:
http://blogs.sun.com/chrisg/page/scsi.d
I have not seen this type of error before. I would think that there might be a
problem with a target and the upper layers send a target reset. Turn on the qlc
driver extended
I know this is starting to get a bit off topic for this forum but you should
look into OS X's two variants of the HFS+ filesystem. There is the normal OS X
filesystem (journaled) which is not case sensitive but is case preserving, and
then there is the OS X filesystem (journaled,
From what I know it should support 4 paths easily. I have a T2000 with 4 HBA
ports in it that I'll hook up to a Xraid later today and check things out.
This message posted from opensolaris.org
___
storage-discuss mailing list
So check if the LUNs are presented using this command:
cfgadm -lva -o show_SCSI_LUN
If all the LUNs are not showing in there, you can check the fcp modules trace:
from a ksh - (echo ::fcptrace | mdb -k) | less
Then look for some lines like this:
[Mon Feb 4 2008 14:01:41:956:635:832]
There is always this free failover solution to try:
http://www.symantec.com/business/theme.jsp?themeid=sfbasic
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
Here are is a proposed change to the page on NPIV:
Before: A WWN is a 64 bit value, and should be world wide unique.
After: A WWN is a 64 bit value, and must be world wide unique.
To quote a gentleman I know speaking about WWPN duplicates and FC switches,
WWPN are required by spec to be
For those curious if the QLogic target mode driver with COMSTAR supports
initiators with MPxIO;
Yes it works. Just follow the directions in this manual:
http://dlc.sun.com/pdf/817-3674-12/817-3674-12.pdf
On my my Solaris 8 initiator host's /kernel/drv/scsi_vhci.conf I changed the
mpxio disable
This is great!
Thank You
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
I bet if you were to bring the configuration back to the two hosts and storage
in the same zone, and then gather the ::fptrace, we could tell what was
happening.
This message posted from opensolaris.org
___
storage-discuss mailing list
Here is a sample of gathering the fptrace with mdb:
root:= mdb -k
Loading modules: [ unix krtld genunix specfs dtrace ufs sd px ip hook neti sctp
arp usba fcp fctl qlc nca lofs md logindmux ptm cpc sppp random crypto fcip zfs
ssd nfs ipc ]
::fptrace
fp trace buffer contents
[Tue Nov 6 2007
I would not say it's hardware, especially if you get a link to the switch.
From the boot failure I see that you are running Solaris 10. If it's U3 or
higher you can gather the output from the layers above the qlc driver. Check
out this excellent presentation:
Just checking because people are extra creative sometimes.
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
From the cfgadm -al output it appears there is no storage available to the
HBA. If there was, it would look something like this:
Ap_Id Type Receptacle Occupant Condition
c0 scsi-bus connectedconfigured unknown
With the newer 4gig HBA release, Sun and QLogic have support for the scli and
SANsurfer utilities to perform some limited operations on qlc driver controlled
HBAs. Now those two utilities use a passthrough IOCTL to communicate with
devices visible through the HBAs.
I would check to see if the
If this is a Solaris 10 host, you could use the scsi.d dtrace script to log the
SCSI commands from this host.
http://blogs.sun.com/chrisg/page/scsi.d
Now you'll end up with a lot of info, and you will need to log the time you
started scsi.d as the timestamps appear to be timestamps from when
I do this:
iostat -xCM 5| egrep -v nfs|sd
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
I just tried it on an E250 with OBP 3.28.0 and it works.
So I'd think being it's worked on two extremes of the current Solaris supported
hardware, it will probably work on what you have.
Lyle
This message posted from opensolaris.org
___
So I just installed U4 on a T2000 and I found it easiest to just delete the HBA
nodes from the OBP before the boot net - install. Then the HBA ports were
never presented to the OS until the machine went through POST again.
for example:
{0} ok /[EMAIL PROTECTED]/[EMAIL PROTECTED]/[EMAIL
No marketing fluff. Just a URL:
http://cvs.opensolaris.org/source/xref/nwsc/src/sun_nws/qlc/
Welcome my son, welcome to the machine
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
I'm excited about this too.
There is a lot of getting up to speed with this release and the release of S10
U4.
This message posted from opensolaris.org
___
storage-discuss mailing list
storage-discuss@opensolaris.org
The leadville stack supports automatic persistent binding by the way it is
designed.
Check out the following docs:
http://www.jmcp.homeunix.com/~jmcp/WhatIsAGuid.pdf
slide 40 touches on the devlinks for the leadville stack
http://developers.sun.com/solaris/articles/fc_drivers_sfs.html
There should NEVER be a reason to modify the /etc/driver_aliases for a
production configuration. If there is trouble with QLogic cards and the qlc
driver not attaching then open a service request on the QLogic website.
With that said. There were some early versions of the FCODE included on the
Gabriele,
The nature of a SCSI check condition is that it is an error detected or
reported by the SCSI target.
Two items support James' conclusion that the tape drive or media are problems.
The first is that the drive reports a Media Error sense key. In SCSI terms,
media is the tape or disk
Tim,
There is no implied auction site or reseller in what I stated.
Here is the official QLogic stance on warranty:
http://support.qlogic.com/support/warranty.asp?id=m12
Transferability. QLogic’s warranty only applies to the original purchaser of
the QLogic product from QLogic (or, in the case
All I really wanted is to switch given fc port off so I do not have to
physically go to data center to remove/plug fc link in order to final
testing if everything is ok.
So my gut response is, That's what rookie sysadmins are for! I didn't spend a
year vacuuming dist bunnies out of sparc-5 hosts
In FC you take a SCSI packet, wrap it into a FC frame (with a CRC at the end),
and then when that FC frame is put out onto the wire/fiber, it is encoded using
this 8b/10b method.
http://en.wikipedia.org/wiki/8B10B
What a invalid Tx word is when there is a 8b/10b decoding error. You'll see
You could install that scli utility of onto one host then NFS share out the
/opt/QLogic_Corporation/SANsurferCLI dir and mount it on all the hosts. Then
just run this command line on each host:
/opt/QLogic_Corporation/SANsurferCL/scli -g
And if the output is not quite what you'd like, then you
46 matches
Mail list logo