We are in the process of moving an apache webserver with windows to
nginx on zLinux.
We use Redis and Memcached already on intel castoff machines.
Management is uncomfortable with this situation and wants us to get
everything on 24/7 supported hardware. My plan was to buy a 1-u intel
server
Berry,
what's the kernel version of your SLES11 SP1 kernel running? I am aware
of a race problem between udev rules and kernel while accessing a sysfs
attribute of a ccwgroup device. It has been fixed with kernel
2.6.32.49-0.3.1. This problem would match your described symptom
udevd-event[554]:
On Tue, 30 Oct 2012 11:01:59 +0200
Roger Evans ro...@autodata.no wrote:
We are in the process of moving an apache webserver with windows to
nginx on zLinux.
We use Redis and Memcached already on intel castoff machines.
Management is uncomfortable with this situation and wants us to get
Hi!
In RedHat V6.2 the setup in /etc/inittab to skip of logon in 3270 does not
work:
1:2345:respawn:/bin/bash
Any hint or tips to make the skip of logon in 3270 to take effect?
Med Vänlig Hälsning / Best Regards
Bertil Starck
Handelsbanken
CDTI-O
tel: +46 8 701 22 51
e-mail:
Hi Ursula,
Thanks. That might be something then. We have 2.6.32.12-0.7 in these guests.
Is there any way to go around this issue? Can we activate the OSA outside udev?
Regards, Berry.
-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of
Ursula
You should do the total cost calculation first.
We usually discover that about half the total cost of a server is
manhours, then there is software(licenses) and monitoring backup
and such things. Cost left for hardware always below 20%, sometimes
quite below.
And surprisingly the x86 platform
On 10/30/2012 at 05:01 AM, Roger Evans ro...@autodata.no wrote:
We are in the process of moving an apache webserver with windows to
nginx on zLinux.
We use Redis and Memcached already on intel castoff machines.
Management is uncomfortable with this situation and wants us to get
everything
On 10/30/2012 at 08:24 AM, van Sleeuwen, Berry
berry.vansleeu...@atos.net
wrote:
Hi Ursula,
Thanks. That might be something then. We have 2.6.32.12-0.7 in these guests.
Is there any way to go around this issue? Can we activate the OSA outside
udev?
Updating the kernel would be the
On RHEL 6, I edit /etc/init/tty.conf and add --autologin root to the
parameters on the mingetty program.
On Tue, Oct 30, 2012 at 6:44 AM, Bertil Starck bes...@handelsbanken.se wrote:
Hi!
In RedHat V6.2 the setup in /etc/inittab to skip of logon in 3270 does not
work:
1:2345:respawn:/bin/bash
But our Z enthusiast
says we can put Redis and memcache on Z-Linux (under VM) without any
loss of functionality or performance (because we have extra capacity and
paging on Z doesn't cost anything).
I'd agree on functionality, but performance is a harder question. Both redis
and memcache
Yeah - we'll probably start small, with say half a GB cache and see if
it affects other packages. We use VDISk for paging, and perfmon shows
that we have XSTOR, but no usage.
I wonder if it would help to put nginx, Memcache and Redis code into a
single VM shared segment. Would it require that
Software support doesn't seem to be an issue for them. We use lots of
open source packages that don't come with any support. Nginx isn't
supported either, but it works fine on Z and we've been using it for
about 18 months without a hitch.
There might be better combinations than Redis/memcache
Thanks Alan - those are my sentiments precisely. And that is what we
have today (except that I don't know for a fact that the memory is ECC).
But sometimes you have to take a bizarre idea and see if there's
something in it. Like Escher is reputed to have said: Only those who
attempt the absurd
On 10/30/2012 at 11:26 AM, Roger Evans ro...@autodata.no wrote:
Software support doesn't seem to be an issue for them. We use lots of
open source packages that don't come with any support. Nginx isn't
supported either, but it works fine on Z and we've been using it for
about 18 months
On 10/30/2012 at 11:46 AM, Roger Evans ro...@autodata.no wrote:
Thanks Alan - those are my sentiments precisely. And that is what we
have today (except that I don't know for a fact that the memory is ECC).
But sometimes you have to take a bizarre idea and see if there's
something in it.
On 10/30/2012 at 11:33 AM, Roger Evans ro...@autodata.no wrote:
Yeah - we'll probably start small, with say half a GB cache and see if
it affects other packages. We use VDISk for paging, and perfmon shows
that we have XSTOR, but no usage.
If z/VM isn't using expanded storage, then the
This may be a bit weird. At present, we are running on a z9BC with two CPs. My
boss is talking with _someone_ about upgrading to a z10 of some sort with 2
CPs. Apparently the vendor is willing to sweeten the deal by throwing in an
IFL. I don't know how powerful an IFL is on a z10. We had one
Don't discount the DASD system on a z box.
My IBM DS6800 manual states that a properly configured DS6800 can support
330,000 I/Os per second. Obviously, no one actually has a properly
configured DS6800 (and I assume that the DS8*** boxes are even better).
My point is that the PC I/O system
On 10/30/2012 at 02:24 PM, McKown, John john.mck...@healthmarkets.com
wrote:
This may be a bit weird. At present, we are running on a z9BC with two CPs.
My boss is talking with _someone_ about upgrading to a z10 of some sort with
2 CPs. Apparently the vendor is willing to sweeten the deal
John, get someone to sit down and use whatever your organization uses for
estimating costs and run the cost-benefit ROI on this before you get started.
On a scale of 1 to 10 where 1 is 'pouring money down a rathole; and 10 is 'we
just saved the farm!' this effort could easily run anywhere
Exactly what doesn't work here ?
We use RHEL6.2 without having to fiddle with this, and
everything work just as we want. We maybe want different things :-)
/BR Tore
Tore Agblad
System programmer, Volvo IT certified IT Architect
Volvo Information
On Tue, Oct 30, 2012 at 11:38:41AM -0600, Mark Post wrote:
OK, this is starting to sound like System z could be a good choice. If the
VSE systems or DB2/MySQL systems are guests of z/VM on the same CEC as the
Linux system would be, you could get very low latency network connections
between
Hi Berry,
sure, you can always activate a qeth device manually. One possiblitiy is
to make use of the znetconf command. Determine your configured options
for your OSA (for instance from the corresponding udev-rule
in /etc/udev/rules.d/51-qeth...) and specify them explicitly with the
-o option of
23 matches
Mail list logo
