In 2 past lives I did a lot of Java tuning. 99.993% of the time the problem was
database queries.
How busy are the JVM's?
How busy is the database server?
How big is the network pipe into the VM's?
What other symptoms do you have?
Edward Long
From: "Smith, A
Meyer, Kenneth J wrote:
> John, you are right on about Verizon. They are no longer the same company.
> I (personal) have ceased our internet/FIOS contract and will be shortly doing
> the same
> with our cell phone service. They forgot what it means to strive for
> excellent
> service that breed
Thanks for the info Mark.
I did find that lszfcp provided info I needed.
lszfcp -a -b 0.0.4800 seems to show local port_name
lszfcp -P -H -D -V has some remote side info
Am I correct that zfcp_host_configure and zfcp_disk_configure are still alive
and well?
They may just update files in /etc/
Well, it kind of comes down to figuring out what resource is the bottleneck.
>From the VM perspective, I'm sure you've looked at cpu wait, page wait, i/o
>wait.
Avoid paging the heap if at all possible...
>From the Linux perspective, any swapping?
>From the Java perspective you often need some so
>>> On 8/25/2013 at 12:23 PM, "Smith, Ann (CTO Service Delivery)"
wrote:
> Upgraded server using EMC san from sles10 sp4 to sles11 sp2
>
> LVM cannot find volume groups
>
> lsscsi shows devices
>
> Cannot find udevinfo command that used to be in udev package or san_disc
> command that used to
Interesting enough I have the same problem but not running Linux-390, but
running it under USS
We are running JVM
# java -version
java version "1.6.0"
Java(TM) SE Runtime Environment (build pmz6460_26sr6-20130619_01(SR6))
IBM J9 VM (build 2.6, JRE 1.6.0 z/OS s390x-64 20130613_152190 (JIT
enabled
I forgot to also add that there is stuff in the Diagnostics Guide
http://publib.boulder.ibm.com/infocenter/javasdk/v6r0/index.jsp that might help.
Marcy
--
For LINUX-390 subscribe / signoff / archive access instructions,
send em
>>> On 9/3/2013 at 01:35 PM, Martha McConaghy wrote:
> We have embarked on a project to support our main student/faculty/staff
> portal
> on zLinux. It works fine under a normal workload, but falls over badly when
> under a heavy load such as during class add/drop. The portal is running open
>
What java policy are you using?
We use java gc 'gencon' policy for incremental java heap for some applications.
Garbage collection may occur more often - but less impact to application when
garbage collection does occur.
Don't know that that is your issue - but something to try.
Annie Smith
--
We have embarked on a project to support our main student/faculty/staff portal
on zLinux. It works fine under a normal workload, but falls over badly when
under a heavy load such as during class add/drop. The portal is running open
source software named LifeRay on top of TomCat on SLES 11 SP1.
We
John, you are right on about Verizon. They are no longer the same company.
I (personal) have ceased our internet/FIOS contract and will be shortly doing
the same
with our cell phone service. They forgot what it means to strive for excellent
service that breeds customer loyalty. I would never tr
JES2 fully supports NJE over TCP/IP and, in addition, supports SSL
connections.
On 9/3/13 9:16 AM, "Rick Troth" wrote:
> NJE can stuff files directly into MVS spool space. This is probably
> the most "normal" way to get files into MVS. You might need to hop
> through VM, since I don't know th
Time out. Back up to the 10 kilometer altitude and consider what
you're trying to do. FTP might not be the right tool. (Ubiquitous,
yes. But ... something about hammers and nails should go here.)
Sounds like the credentials fight is your big pain point. The
credentials requirement is a big tu
If you bill your customer by consumed cputime, you probably get these figures
from z/VM
So in that case you probably want cpustatistic from z/VM as well, so they match.
/Tore
-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Angel
Tamayo
Sent: den 3
14 matches
Mail list logo