Greetings; (Posted to VMESA-L and VSE-L and LINUX-390)
- - Now in its fifth year! - - Now includes VSE and linux/390!
I have set up a public service web page at
http://www.eskimo.com/~wix/vm/
for posting positions available and wanted for VM, VSE and linux/390.
Please visit the
Asher Glynn writes:
Has anyone read a book on the Linux kernel that they would recommend buying?
Depending on what parts of the kernel you're interested in,
Linux Device Drivers (Rubini) and Understanding the Linux Kernel
(Bovet Cesati), are worth reading. Both published by O'Reilly.
There
Has anyone read a book on the Linux kernel that they would recommend buying?
Thanks,
Give Mr Molnar's a miss.
--
Cheers
John Summerfield
Microsoft's most solid OS: http://www.geocities.com/rcwoolley/
Note: mail delivered to me is deemed to be intended for me, for my disposition.
The fatal flaw in META'a arguement is that Linux or no, mainframes are not
going anywhere soon. Mainframes have been appearing in the obituary
section of IT mags for quite a long time now, yet inexplicably have yet to
be totally replaced by PCs :)
The current loss rate has a number of people
- Understanding the Linux Kernel - Bovey Cesati - O'Reilly
- The Linux Kernel Book - Card, Dumas, Mevel - Wiley
- Linux - Michael Kofler - Addison Wesley
- Linux Device Drivers - Rubin - O'Reilly
-Original Message-
Has anyone read a book on the Linux kernel that they would
recommend
Hi,
You can use the logger command on Linux/390 and point the /etc/syslogd.conf
at your VM host. You can then install my TCPLOGD server (on the 1997 VM Tools
Tape) and customize it to your heart's content. The TCPLOGD server is a
REXX EXEC and a NAMES file. That's it. Piece of cake.
Hello Mark, thanks for your help,
I changed ctc0 to escon0 in my chandev.conf without success.
I changed my iodf has recommended.
I tried protocol_no value 0, 1, 3 without success.
When I try protocol_no=2 i get No such a device even after create
/dev/ttyZ0.
This is what I get with cat
Based on my install experience, that sounds reasonable. There must have been
a time when the JDBC levels were mixed between the products. That seems to
be behind us at this point.
On Thursday 16 May 2002 07:51 am, you wrote:
I opened an incident with IBM. This is what they said. The
The META report says that mainframe configurations of Linux will fall out
of favor as soon as 2005, by which time Unix and Intel-based Windows and
Linux solutions will have enough mainframe-like features without the
mainframe-like costs to make those options the better choice. In a scathing
Applying linux-2.4.17-s390.diff by itself and building the kernel results in dasd.c
compile failing, so it appears the 2.4.17 source tree on developerworks is seriously
broken...
Patrick A. Carroll
Senior Systems Engineer
L. L. Bean
Casco St
Freeport, ME 04033
(207) 552-2426
[EMAIL
I have heard that multipath i/o is supported on s390 and was curious how
this is accomplished. If anyone has any information I'd appreciate a
reference or a reply to this post.
On other architectures, multipath i/o is handled in the scsi layer, using
the scsi id found on newer drives.
Historically Meta have always been very anti Linux ...
http://www.itworld.com/nl/it_insights/09042001/pf_index.html ?
--
Phil Payne
http://www.isham-research.com
+44 7785 302 803
+49 173 6242039
Thanks to all who replied to this question. Looks like everyone who replied did
so directly to me instead of the forum, so I'll include this for reference.
This reply from Mark is what did the trick. We are using the Sistina LVM and the
Howto is short and to the point.
The only problems I
On Thu, May 16, 2002 at 03:51:25PM +0200, Phil Payne wrote:
Historically Meta have always been very anti Linux ...
http://www.itworld.com/nl/it_insights/09042001/pf_index.html ?
This article looks to me like it's damning with faint praise: they saw the
handwriting on the wall, and had to say
Good Morning -
I am trying to install redhat 7.2 linux on a s390 lpar. I downloaded
the tapeinrd.img and the tapekrnl.img from ftp.redhat.com in
/pub/redhat/linux/7.2/en/s390/images and created a parmfile. The system
Boots fine and I get the prompt for a FQDN, etc. and a device type which
if I
I have heard that multipath i/o is supported on s390 and was curious
how this is accomplished. If anyone has any information I'd appreciate
a reference or a reply to this post.
On other architectures, multipath i/o is handled in the scsi layer,
using the scsi id found on newer drives.
Multipathing is handled by the I/O processors of the S/390. Linux doesn't
even know (or care) that it is happening.
On Thursday 16 May 2002 08:45 am, you wrote:
I have heard that multipath i/o is supported on s390 and was curious how
this is accomplished. If anyone has any information I'd
Depends on your need, of course. There's the O'Reilly book
Understanding the Linux Kernel
(http://www.oreilly.com/catalog/linuxkernel), but it's meant as an
OS course text or supplement, and is a bit dated. It focuses on the
2.2 kernel (with sections at the end of each chapter about changes
to
On Thu, 16 May 2002, Don Mulvey wrote:
I have heard that multipath i/o is supported on s390 and was curious how
this is accomplished. If anyone has any information I'd appreciate a
reference or a reply to this post.
On other architectures, multipath i/o is handled in the scsi layer,
In the redbook Linux for s/390 and zSeries: distributions there is a
chapter that shows how to setup LDAP for authentication. Authenticating
against RACF has limitations because you can only have one RACF user
profile with a specific name. In other words one ROOT user only. SO if you
are doing a
In the case of connecting SCSI devices to the Shark, the system to be
connected has two SCSI cards and each card is connect to separate SCSI
adapters on the Shark. There is also a special driver that needs to be
installed (called DPO, dual path option) that handles the multipathing
function.
On
Barton Robinson [EMAIL PROTECTED] writes:
Tom Beretvas, Mr. DASD, previously of IBM, author of many
performance papers and presentations on DASD performance. Tom is
bringing his world class DASD performance background to the
Linux and VM environment to help us all better understand how to
I looked at the article. They seem to have a single-minded view that Linux
is going to, actually _must_, fragment or fork at some point in the
future. That colors the rest of their analysis for me, as it seems to
indicate a severe lack of understanding of the Linux community. There have
always
necessarily leading to a permanent forking of the code. Even if it does,
Linus owns the rights to the Linux trademark, and if someone goes too far
overboard, he has the ability to prevent them from calling it Linux. I
would be surprised if it ever goes that far.
Whats interesting is that
Whats interesting is that not only do the distributors want to avoid
fragmentation (your maintenance cost rises steeply with each shift you make
away from the base tree) but the customers understand the importance.
The enterprise customers are quite open that anyone who goes off on a weird
Here is the URL for the article Mark referred to.
http://vb.channelsupersearch.com/news/var/35257.asp
Salomon Smith Barney is a unit of CitiGroup.
Regards, Jim
the OS doesn't need to know about the multiple paths
For high availability, yes. But for performance, I was *under the
impression* that Linux needs to be fooled into using the multiple
paths (haven't been able to confirm this with end-to-end performance
tests). This is done by LVM or
For anyone that ordered the free IBM Linux SEK, it has been updated. I just
got this email notification today. Please note that this is for Intel Linux
systems only.
When you downloaded the IBM Software Evaluation Kit (SEK) for
Linux, you requested to be notified by e-mail when IBM updates
It turns out there's a PDF version of the whole LVM HOWTO at
http://www.sistina.com/lvm_howtos/lvm_howto.pdf
Mark Post
-Original Message-
From: p43cibmgs-Arden [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 16, 2002 10:13 AM
To: [EMAIL PROTECTED]
Subject: Re: Anyone know how to take
Hi,
With the activities I have seen on this mailing list, I am sure most of you
are experts at cloning Linux images/instances on z/VM by now. For some of
the new people who just got started with Linux on zSeries/S390, I thought
it may be still be helpful to share my tools/experience on
Steve Arden wrote:
Does anyone know how to move Logical Volumes from one Linux guest to
another?
I haven't tried 2.2.16 to 2.4.7, but I have moved LVM between LPAR systems.
1) on the new system, do a vgscan to locate all the LVM volumes
2) activate the volume groups with vgchange -a y vgname
A unit of, but not the whole group. Doctor Robinson from Salomon Smith
Barney gave a presentation at SHARE 98 on his group's use of Linux/390 that
was very well attended. But I don't consider the use of Linux/390 by a
subsidiary equivalent to the holding company adopting Linux/390. Terrance
Under 2.4.7, you can use the sysctl command to interrogate and set many
kernel variables, including max files
sysctl -a | less to see the variables
sysctl -w fs.file-max=16834 to increase the maximum number
Regards, Jim
Linux S/390-zSeries Support, SEEL, IBM Silicon Valley Labs
On Thu, 16 May 2002, Carsten Otte wrote:
Hi List-Readers!
The kernel of current Linux-Distributions does not support
muliple pathes to a dasd device at all.
A workaround is to spread the data over multiple devices
using LVM or MD in striping mode. Using the same amount
of devices like the
This is a typical issue with Linux guests. Linux likes to cache it's file
system in memory and access of that cache is what you are seeing. There
doesn't seem to be a way to turn that function off, but it's ramifications
can be reduced to a certain extent.
Purposely constraining Linux's memory
Lionel:
Without the timer patch, Linux wakes up every 100 milliseconds or so.
That causes VM to assume it's actually doing something, so its working
set never gets trimmed. Your best bet is to apply the timer patch.
Romney
On Thu, 16 May 2002 09:51:18 -0700 Lionel Dyck said:
Is there any
Well, there seems to be some disagreement here.
And some misunderstanding as well. When Jim talks about
multiple paths for the I/O, it is multiple paths to the
control unit (that has a bunch of devices). So you can
have multiple I/O's active at the same time to different
devices on the same
At 19:47 16-05-02, Romney White wrote:
Without the timer patch, Linux wakes up every 100 milliseconds or so.
That causes VM to assume it's actually doing something, so its working
set never gets trimmed. Your best bet is to apply the timer patch.
But 'never' as in 'unless it really has to' If
Just for the sake of accuracy, it's 100Hz, or every 10 milliseconds, a whole
order of magnitude worse.
Mark Post
-Original Message-
From: Romney White [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 16, 2002 1:47 PM
To: [EMAIL PROTECTED]
Subject: Re: High working set
Lionel:
Without
Well, there seems to be some disagreement here.
And some misunderstanding as well. When Jim talks about multiple paths
for the I/O, it is multiple paths to the control unit (that has a
bunch of devices). So you can have multiple I/O's active at the same
time to different devices on the same
Microsoft's claim to the word Windows suffered another blow this week
when a federal judge again questioned the company's assertion that the term
is not generic.
The judge also repeated his denial of Microsoft's request to shut down the
Lindows.com site and block its owner from advertising its
Hello-
I have an interesting issue (at least I think it's interesting).
I'm running redhat 7.2 with the 2.4.9-21 kernel, I pulled down the OCO
modules for the qeth and qdio drivers from developerworks (for my kernel
version), and attempted to load them into the kernel (after the
Tom,
As you've discovered, there are some oddities about the way ESCON channel
definitions work with Linux 2.4. One of those is that you should pass a
value of escon0 to the chandev layer, but the actual device name that gets
created is ctc0. I view this as a bug, but haven't pursued it, yet.
This could be a mis-match between the kernel and module in terms of whether
module versioning is turned on or not, or whether SMP is turned on or not.
Try doing
grep remove_wait_queue_R /proc/ksyms
and see what comes out. On my Red Hat 7.2 system, I get the matching
0001d4b4
From what I've seen I think it is typically set to 256K.
On Thursday 16 May 2002 03:13 pm, you wrote:
I am looking for a parameter that I can set on my Linux system for the
Maximum Network I/O Buffer Size (a parameter that controls the size of the
buffer used to receive blocks of data from
I would probably go for the April 15th one, but that may mean you'll need to
put on a bunch of other patches as well:
linux-2.4.7.tar.gz (original kernel source)
+ linux-2.4.7-s390.diff (IBM)
+ linux-2.4.7-s390-1.diff (IBM)
+
Mark - thanks. There was a collision with the iucv module with patch 7
which prompted my question.
Lionel B. Dyck, Systems Software Lead
Kaiser Permanente Information Technology
25 N. Via Monte Ave
Walnut Creek, Ca 94598
Thanks for the suggestions - I did make sure the my read and write
channels cross connected. I was reading the Device Drivers and
Installation Commands for Linux Kernel 2.4 for S/390 and some things
that I don't understand are:
What is the difference between CTC/ESCON with the channel device
such as stock traders to retrieve and digest
financial statistics and news. The software runs on both workstations and
servers.
http://story.news.yahoo.com/news?tmpl=storyu=/cn/20020516/tc_cn/linux_grabs
_big_win_with_reuters
Mark Post
Tom,
I had similar problems going from RH6.2 to 7.2. I also have the same
configuration you are using (ESCON CTC, Linux LPAR, and z/OS 1.2 on the
other end). In RH6.2 I would define escon0=, but had no success with
that in 7.2. Here is what I did get to work (with a lot of help from
Mark and
This is where distributors make their money. I spent a week configuring and
recompiling theSLES7 kernel to include NO-100Hz-Support, only to find that the OCO's
supplied by IBM don't support SMP. IBM instructed me to get the RPM's from SuSE.
They installed in about 2 minutes.
Cross-posting to VMESA-L, LINUX-390, and IBM-MAIN.
Enroll today to reserve your seat for zSeries education:
IBM eServer zSeries Technical Conference
... featuring z/OS, z/VM, and Linux on zSeries
June 24-28, 2002
IBM International Education Center
La Hulpe, Belgium
Visit the
I am installing Red Hat 7.2 in an LPAR and am using Anaconda. I am at the
stage where I am partitioning the DASD with fdasd. I used command n to
create the new partitions and when I issue command p, I see the
partitions. I then use command w to exit but I find that the partitioning
is lost.
53 matches
Mail list logo
