Hi All,
After we have extended disks in an Oracle server the RMAN backup is killed by
the OOM killer. The full archive runs fine (ie, low memory usage) but the
incremental immediately pushes all allocated memory into swap and then the OOM
killer kicks in.
The guest is running 64bit SLES11 SP1
We are in search of a good VM/Linux Monitor. Until recently, we had Velocity
and were quite happy with it. Barton's group decided our shop was too small
and not worth the hoops that our purchasing division wanted them to jump
through, so now we need a new one. Any ideas?
The guest has 6G memory and 4G swap.
...
How can we tune this system to avoid getting hit by the OOM condition?
Did you try giving the virtual machine more memory? Is 6G a special
number for this workload?
Mike MacIsaac mikemac at-sign us.ibm.com
I have a vested interest as I work for IBM, but OMEGAMON for z/VM and
Linux for
System z is an option..
http://www.ibm.com/software/tivoli/products/omegamon-xe-zvm-linux/index.html
Cheers, Dave S.
IBM United Kingdom Limited
Senior I.T. Specialist - System z Tivoli Technical
Gene,
Be aware that most monitoring is not quite suited for virtual machines. They
don't (or can't) match linux values with the host numbers.
You might want to look at Performance Toolkit or Tivoli Omegamon. Obviously
most monitoring from general linux can also monitor linux on z. Such as
If anything I even would like to decrease the memory configuration rather than
increase it. Unfortunately I got overruled by the Oracle guys who still think
more memory means better performance. During the migration from Oracle 10 to
Oracle 11 we have gone up from 5400M to 6G. I don't like to
IBM announces a $1B investment in Linux on System p, while the story is on
Linux on System z (I guess z already got *its* $1B :)) ...
https://www.linux.com/news/enterprise/systems-management/740850-demand-for-linux-on-system-z-accelerates-ibm-clients-continue-to-see-cost-savings-
Mike MacIsaac
Really good performance monitors would give you lots of good information.
On 10/2/2013 4:28 AM, van Sleeuwen, Berry wrote:
If anything I even would like to decrease the memory configuration rather than
increase it. Unfortunately I got overruled by the Oracle guys who still think
more memory
True, knowing what buttons to push and what dials to turn might help too :-).
We know the memory and swap are exhausted, and we know the rman processes are
the ones that cause it. So now we have to resolve that. I would like to have
rman use less memory but I'm not quite sure if and how we can
What I like about Performance Toolkit for z/VM (Perfkit) besides working
on and supporting the product while at the lab in Endicott... is that's
the tool the z/VM lab still uses helping to simplify collaboration with
the lab when/if required. Understand that it is a hardware/system tool
over
The kernel version that you are using supports cgroups -
https://www.suse.com/releasenotes/s390x/SUSE-SLES/11-SP1/#ResourceManagement
and
has it enabled by default (if the memory cgroups feature is not needed, it
can be switched off by passing cgroup_disable=memory on the kernel command
line,
On Wednesday, 10/02/2013 at 06:39 EDT, Walters, Gene P
gene.p.walt...@wv.gov wrote:
We are in search of a good VM/Linux Monitor. Until recently, we had
Velocity
and were quite happy with it. Barton's group decided our shop was too
small
and not worth the hoops that our purchasing division
If you can't get Velocity's suite of tools (they're excellent), try
taking a look at Perfkit and OmgeaMon XE for zLinux. OmegaMon XE for
zLinux can also feed performance data into a z/OS - OmegaMon
environment, if you're running that as well.
DJ
On 10/02/2013 05:37 AM, Walters, Gene P wrote:
We
Note that CA can also sell you licenses for at least some of Velocity's
products, so if you have a relationship with CA, that might be an easier route
to go than making your procurement guys do anything out of the ordinary.
From: Linux on 390 Port
Hi, All,
We're having fun with installing a vendor's web server into a RHEL 6.3 guest
on z/VM 6.2. Vendor's support person talked us through setting the SELinux
configuration to disabled and rebooting Linux, but it still comes up with
SELinux showing permissive. The installation appears to
You need change /etc/selinux/config to disabled.
# This file controls the state of SELinux on the system. # SELINUX= can
take one of these three values: # enforcing - SELinux security policy is
enforced. # permissive - SELinux prints warnings instead of enforcing. #
disabled - No SELinux policy
On 10/02/2013 12:37 PM, Chase, John wrote:
We're having fun with installing a vendor's web server into a RHEL 6.3
guest on z/VM 6.2. Vendor's support person talked us through setting the
SELinux configuration to disabled and rebooting Linux, but it still comes
up with SELinux showing
John, what does the sestatus command display?
DJ
On 10/02/2013 11:37 AM, Chase, John wrote:
Hi, All,
We're having fun with installing a vendor's web server into a RHEL 6.3
guest on z/VM 6.2. Vendor's support person talked us through setting the
SELinux configuration to disabled and
I heard that activating DIRECT_IO helps solve the RMAN memory eating
problem on kernels 2.4.
On http://dbasolutions.wikispaces.com/RMAN+Performance+Tuning you can see
some ideas on how to tune rman memory usage.
To show your Oracle DBA's that more memory is less performance, ask them to
try the
Did you try setting SELINUX=permissive in /etc/selinux/config?
Mike MacIsaac mikemac at-sign us.ibm.com
--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message:
On Wed, 2 Oct 2013, Chase, John wrote:
What else must be done to get the RHEL vm to come up with
SELinux showing disabled instead of permissive in
response to a 'getenforce' command?
adding a line to the relevant grub stanza comes to mind.
Something like this
title SE-Linux Test System
Or just specify the parm selinux=0 on the IPL command
DJ
On 10/02/2013 11:44 AM, Rick Troth wrote:
On 10/02/2013 12:37 PM, Chase, John wrote:
We're having fun with installing a vendor's web server into a RHEL 6.3
guest on z/VM 6.2. Vendor's support person talked us through setting the
On Wed, 2 Oct 2013, R P Herrold wrote:
kernel /boot/vmlinuz-2.4.20-selinux-2003040709 ro
root=/dev/hda1 nousb enforcing=0
as:
kernel /boot/vmlinuz-2.4.20-selinux-2003040709 ro \
root=/dev/hda1 nousb enforcing=0
(no backslashes are handled by grub, and the root device
Do you really want to handicap the security on your Linux server by disabling
SELinux? I use the audit2allow command as outlined at
http://www.linuxforums.org/articles/accomodating-avc-denied-messages-selinux_355.html
to create and load needed local policies for SELinux. It is an iterative
On 10/2/2013 at 12:52 PM, R P Herrold herr...@owlriver.com wrote:
adding a line to the relevant grub stanza comes to mind.
I'm not aware of grub working on System z (yet).
Mark Post
--
For LINUX-390 subscribe / signoff /
RHEL 5.8 zLinux on zVM 6.1 using ECKD disks
We have a recurring problem where files get corrupted. It always happens in
the same directory, and we always wind up running fsck in single user mode
to get the file system back together.
The file system is EXT-2. (We changed from EXT-3 on this FS due
I ran RHEL 5.x for many years on z/VM 5.4 and on z/VM 6.2 without any file
corruptions. Have you verified that there is not a minidisk overlay on z/VM
side? Of course a minidisk overlay would probably corruption move than one 4K
block.
-Original Message-
From: Linux on 390 Port
Data corruption should be *very* rare in any working system. The majority
of the cases occurs when two systems write on the same disk, uncoordinated.
When you changed from ext3 to ext2, you threw away the only thing
preventing this to happen. Ext3 detected something amiss, and locked the
disk to
I also would be suspicious of mini-disk overlay in ZVM. I did it myself, and
for the most part the Linux instances ran fine.
The information contained in this e-mail message is intended only for the
personal and confidential use of the designated recipient(s)
Donald,
You could check that you are not inadvertently using a file system, lvm or
swap on an entire dasd instead of in a partition. If you are, typical
symptoms could be seeing e5, for example, if a process SIGSEGV'd
Registers:
r0 e5e5e5e5e5e5e5e5 r1 86ddccb8 r2 86dda218
-Original Message-
From: Linux on 390 Port On Behalf Of Hodge, Robert L
Do you really want to handicap the security on your Linux server by disabling
SELinux? I use the
audit2allow command as outlined at
http://www.linuxforums.org/articles/accomodating-avc-denied-
On 10/02/2013 01:00 PM, Hodge, Robert L wrote:
Do you really want to handicap the security on your Linux server by disabling
SELinux? I use the audit2allow command as outlined at
http://www.linuxforums.org/articles/accomodating-avc-denied-messages-selinux_355.html
to create and load needed
In my book, any Linux error that doesn't make sense is a SELinux error. I then
run aureport -a to verify the assumption.
-Original Message-
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Chase,
John
Sent: Wednesday, October 02, 2013 12:22 PM
To:
-Original Message-
From: Linux on 390 Port On Behalf Of Hodge, Robert L
In my book, any Linux error that doesn't make sense is a SELinux error. I
then run aureport -a to
verify the assumption.
[RHEL63]# aureport -a
AVC Report
On Wed, 2 Oct 2013, Chase, John wrote:
Part of the problem is that SELinux is not writing any
messages anywhere we can see
In a Red Hat derived environment, when enabled SELinux uses
one of two possible file locations when in enforcing, or
permissive modes:
/var/log/messages
-or-
No flame war. Strong security is a corporate policy. I don't feel that I have
the option of disabling security. It is just another component of the Linux
system to feed and take care of. The learning curve is not steep once you
recover from the initial shock of it being there. The resource that
[WARNING] This e-mail post is intended to be more humorous than
serious by poking fun at an organization known to have less than a low
sense of humor [/WARNING]
Y'know, the subject line of Disabling SELinux caught my eye... since
the NSA had a big piece of the patches that enhance Linux security.
On 10/2/2013 at 02:43 PM, Chase, John jch...@ussco.com wrote:
[RHEL63]# aureport -a
AVC Report
# date time comm subj syscall class permission obj event
no events of interest
-Original Message-
From: Linux on 390 Port On Behalf Of John Campbell
[WARNING] This e-mail post is intended to be more humorous than serious by
poking fun at an
organization known to have less than a low sense of humor [/WARNING]
Y'know, the subject line of Disabling SELinux
John, I think you will need to install the following:
compat-libstdc++-295
compat-libstdc++-33
and maybe
compat-2006.1.25-11.2.s390x.rpm
compat-32bit-2006.1.25-11.2.s390x.rpm
That should at least get you started..good luck.
DJ
On 10/02/2013 02:41 PM, Chase, John wrote:
-Original
Hi Mauro,
Would that also be the case in 2.6 kernels?
Anyway, we already have direct_io enabled. Back in 2006 we had only async_IO
and hit a kernel bug because of it. After direct_io was activated as well the
kernel bug didn't happen again. (I think the IO load was so high that the async
Thanks Mauro,
I've checked for overlaps, and confirmed there are none that could account
for this my next step is to check the swap files as Michael
suggests though if we had those sorts of problems I'd expect to see
much more corruption.
The investigation continues...
Cheers,
Donald
On 10/2/2013 at 09:29 PM, Donald Russell russell@gmail.com wrote:
I've checked for overlaps, and confirmed there are none that could account
for this my next step is to check the swap files as Michael
suggests though if we had those sorts of problems I'd expect to see
much more
43 matches
Mail list logo