Wow, sounds almost painless. We are going to go that route at some point.
You know, Mark, that enough people in enough places owe you enough beer
that you can probably go anywhere on earth and not want for a drink.
|---------+---------------------------->
| | "Post, Mark K" |
| | <[EMAIL PROTECTED]|
| | m> |
| | Sent by: Linux on|
| | 390 Port |
| | <[EMAIL PROTECTED]|
| | IST.EDU> |
| | |
| | |
| | 09/24/2003 09:42 |
| | AM |
| | Please respond to|
| | Linux on 390 Port|
| | |
|---------+---------------------------->
>------------------------------------------------------------------------------------------------------------------------------|
|
|
| To: [EMAIL PROTECTED]
|
| cc:
|
| Subject:
|
>------------------------------------------------------------------------------------------------------------------------------|
You can't go to FCP completely. Your IPL volume still needs to be on DASD
(at least for now). For the rest, yes, just move the data, mount the new
file system, update /etc/fstab, remove the old volume from your parmfile,
re-run zipl, do _not_ re-IPL, echo "set device range=xxx-yyy off " >
/proc/dasd/devices, and you're done. (I may have missed something there,
since this is just off the top of my head, but it's close.)
Mark Post
-----Original Message-----
From: James Melin [mailto:[EMAIL PROTECTED]
Sent: Wednesday, September 24, 2003 10:34 AM
To: [EMAIL PROTECTED]
Subject:
Lets say you go and built the right environment to support FC SCSI directly
to the shark box.....but you have these CDL volumes lying around because of
where you were previously...
IS it just a matter of moving the file systems off of ECKD emulated dasd,
to the SCSI attatched disk and changing the IPL parameters or what?
Basically how do you from 3390 emulated devices to completely onto SCSI
over FC.
|---------+---------------------------->
| | "Post, Mark K" |
| | <[EMAIL PROTECTED]|
| | m> |
| | Sent by: Linux on|
| | 390 Port |
| | <[EMAIL PROTECTED]|
| | IST.EDU> |
| | |
| | |
| | 09/24/2003 09:24 |
| | AM |
| | Please respond to|
| | Linux on 390 Port|
| | |
|---------+---------------------------->
>---------------------------------------------------------------------------
---------------------------------------------------|
|
|
| To: [EMAIL PROTECTED]
|
| cc:
|
| Subject:
|
>---------------------------------------------------------------------------
---------------------------------------------------|
I'm not exactly sure what you mean by "is there a conversion strategy." I
need more context to answer it.
I know for sure that the 2.4.19 kernel with IBM's patches has support for
FCP. I cannot say if the 2.4.17 patches had it or not.
$ locate zfcp
/usr/src/linux-2.4.19/include/config/zfcp.h
/usr/src/linux-2.4.19/include/config/zfcp
/usr/src/linux-2.4.19/include/config/zfcp/module.h
/usr/src/linux-2.4.19/drivers/s390/scsi/zfcp.c
/usr/src/linux-2.4.19/drivers/s390/scsi/.zfcp.o.flags
/usr/src/linux-2.4.19/drivers/s390/scsi/zfcp.o
/lib/modules/2.4.19/kernel/drivers/s390/scsi/zfcp.o
Mark Post
-----Original Message-----
From: James Melin [mailto:[EMAIL PROTECTED]
Sent: Wednesday, September 24, 2003 8:30 AM
To: [EMAIL PROTECTED]
Subject:
Mark, a corollary question on point #3 here
Is there a conversion strategy to going this route? If so what kernel level
do you need to be at etc?
|---------+---------------------------->
| | "Post, Mark K" |
| | <[EMAIL PROTECTED]|
| | m> |
| | Sent by: Linux on|
| | 390 Port |
| | <[EMAIL PROTECTED]|
| | IST.EDU> |
| | |
| | |
| | 09/23/2003 04:01 |
| | PM |
| | Please respond to|
| | Linux on 390 Port|
| | |
|---------+---------------------------->
>---------------------------------------------------------------------------
---------------------------------------------------|
|
|
| To: [EMAIL PROTECTED]
|
| cc:
|
| Subject:
|
>---------------------------------------------------------------------------
---------------------------------------------------|
--- snip --
John,
3 - It's already been done with FCP support. Systems can access storage
arrays "natively" without going through that [EMAIL PROTECTED] 3390-emulation layer any
longer. (I'm assuming you meant 3390 and not 3380 as you typed it.) Any
distribution can support that many 3390 volumes, though. Whether the
distribution provider has that support "as distributed" or whether the IT
shop has to create the device nodes is another question.
M
---snip ---
