David Boyes wrote:
since we're running the CPU at 100%, we don't think the problem is all
page waits, although there have been some kernel messages from the
Websphere machine in particular that memory allocations for Java
processes (tasks?) are failing.
Check the max heap size for the JVM.
Ferguson, Neale wrote:
Yes, but if you're not seeing anything in E3 then I guess those settings aren't problematic (although they are set for a more CMS intensive environment).
They're basically unmodified. I haven't had to tune them since we IPL'ed
to remove the V=R region.
Do you use QUICKDSP?
Nick Laflamme wrote:
Well, it didn't take long for the problem to reappear this morning
So you're spending time in system and not in userspace. The memory
numbers look very small, and you have not swapped a dime? Are you sure
you *have* swap space defined? You were showing pictures of a
Rob van der Heij wrote:
Nick Laflamme wrote:
Well, it didn't take long for the problem to reappear this morning
So you're spending time in system and not in userspace. The memory
numbers look very small, and you have not swapped a dime? Are you sure
you *have* swap space defined? You were
You can tell for sure with the 'free' command.
On Fri, 2004-03-26 at 07:54, Nick Laflamme wrote:
Rob van der Heij wrote:
Nick Laflamme wrote:
Well, it didn't take long for the problem to reappear this morning
So you're spending time in system and not in userspace. The memory
On Fri, 2004-03-26 at 07:14, Nick Laflamme wrote:
Do I replace the file system with one that's not ReiserFS, or what? Is
the JVM max heap size somehow tunable?
You might try swapping ext3 for ReiserFS. Are you watching the
console? If ReiserFS is blowing up you will see a kernel Oops on the
What do your swap rate and vmstat output look like?
was1:~ # vmstat
[... vmstat output ...]
[... problem description ...]
That vmstat looks OK. From the problem description it sounds more like
it's JVM related than file system though. When the JVM initializes, it
mallocs and zeros its
Laflamme [SMTP:[EMAIL PROTECTED]
Sent: Thursday, March 25, 2004 3:36 PM
To: [EMAIL PROTECTED]
Subject: Processor starvation
We're trying to work with Websphere and Oracle 9i as Linux guests (one
for each) on an MP 3000 H30 running z/VM 4.4 as a hypervisor.
Admittedly, this is a small box (1
Is
the JVM max heap size somehow tunable?
Yes, but I need to look up the syntax. What version of the JRE are you
using?
I don't know where you would set this in WAS, but the java commandline
arguments are
-Xmx (Max Size)
-Xms (Starting Size)
---
Jeremy Warren
We're trying to work with Websphere and Oracle 9i as Linux guests (one
for each) on an MP 3000 H30 running z/VM 4.4 as a hypervisor.
Admittedly, this is a small box (1 GB of storage, and one CPU that's not
even running at full-power for an MP 3000), but we're seeing horrible
inconsistent response
What are your SRM settings?
-Original Message-
Most disturbing have been a couple of episodes in which the Websphere
guest just simply stops responding to terminal inputs. Our terminal
sessions are via ssh, but these sessions seem to just hang. New
connections time out, and logged in
[EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Thursday, March 25, 2004 3:36 PM
Subject: Processor starvation
We're trying to work with Websphere and Oracle 9i as Linux guests (one
for each) on an MP 3000 H30 running z/VM 4.4 as a hypervisor.
Admittedly, this is a small box (1 GB of storage, and one
Ferguson, Neale wrote:
What are your SRM settings?
IABIAS : INTENSITY=90%; DURATION=2
LDUBUF : Q1=100% Q2=75% Q3=60%
STORBUF: Q1=125% Q2=105% Q3=95%
DSPBUF : Q1=32767 Q2=32767 Q3=32767
DISPATCHING MINOR TIMESLICE = 5 MS
MAXWSS : LIMIT=%.. : PAGES=99
XSTORE : 0%
Do SRM settings
Paging rates are sometimes several hundred pages a second to both
e-store and disk. Ugly, but true, and adding the volume seemed to remove
a bottleneck that had kept us from maxing out the CPU.
Rich Smrcina wrote:
You indicate that you are adding paging volumes. What is VM's paging
rate when
Rob Schwartz wrote:
Nick,
We had the same episodes with un-responsive guests. These virtual
machines ended up on the E3 list. We've adjusted our LDUBUF and STORBUF as
show below. This allows users not to get stuck on E3. You can see if
users are on E3 by doing the indicate load command.
Yes, but if you're not seeing anything in E3 then I guess those settings
aren't problematic (although they are set for a more CMS intensive
environment). Do you use QUICKDSP? What else is running? What does IND Q and
IND Q EXP report?
-Original Message-
IABIAS : INTENSITY=90%; DURATION=2
since we're running the CPU at 100%, we don't think the problem is all
page waits, although there have been some kernel messages from the
Websphere machine in particular that memory allocations for Java
processes (tasks?) are failing.
Check the max heap size for the JVM. If you're running
17 matches
Mail list logo