01-02-2013
The load balancer is going to be a choking point (bandwidth and CPU utilization-wise). If it's bandwidth you may consider increasing the number of links. You can try to increase links through interface bonding. If it's CPU utilization, you might consider multiple load balances (kind of like google.com does) so you can spread load amongst several physical machines. I'd recommend doing that latter anyways, for availability reasons.
Before anyone can give you specific advice though, you need to localize the performance issue. For example, if you run a jMeter test against the serving nodes directly is that faster? If it is, then the load balancer is the choke point. If not then you may check to see if the network links are being saturated with some sort of command-line bandwidth checking tool (I use bwm-ng). If that's not the issue then move onto CPU, memory, etc. Once you've done that you should be in a better position to do something about the performance problems.
It's also possible that it's an application-level latency. For example if a particular website is served by a particular DB cluster and that cluster is going slow it may slow down the end user's experience.
10 More Discussions You Might Find Interesting
1. High Performance Computing
Hi ! I have a simple setup of 2 PC (with linux Red-Hat) where the first PC is the primary machine and the second the backup. I use DRBD for data replication and Red-Hat cluster suite for HA (High Availability). I have tested both.
Now I NEED a COMMON IP ADDRESS (or Master/unique IP address) for... (0 Replies)
Discussion started by: Danny Gilbert
0 Replies
2. UNIX for Dummies Questions & Answers
Hi ! I have a simple setup of 2 PC (with linux Red-Hat) where the first PC is the primary machine and the second the backup. I use DRBD for data replication and Red-Hat cluster suite for HA (High Availability). I have tested both.
Now I NEED a COMMON IP ADDRESS (or Master/unique IP address) for... (3 Replies)
Discussion started by: Danny Gilbert
3 Replies
3. Emergency UNIX and Linux Support
hello Gurus,
My current set up is 3 to 1 Cluster (SUN Cluster 3.2) running oracle database. Task is to reboot the servers. My query is about the procedure to do the same.
My understanding is suspend the databases to avoid switchover. Then execute the command scshutdown to down the cluster... (4 Replies)
Discussion started by: EmbedUX
4 Replies
4. Red Hat
Hi all, i have 3 nodes cluster (Centos 5 cluster suit) with out quorum disk,
node vote = 1,
the value of a quorum = 2,
when 2 nodes going offline, cluster services are destoys.
How i can save the cluster and all services(move all services to one alive node)
with out quorum disk when other... (3 Replies)
Discussion started by: Flomaster
3 Replies
5. Solaris
I am new to setting up sun solaris 10 cluster, I have 2 sun sparc t3-1 servers (identical), going to use them as web servers (sun one java web server 7), looking for data replication and real time fail over. My question is do I need external storage to configure the cluster? or I can just use... (3 Replies)
Discussion started by: spitfire2011
3 Replies
6. Solaris
Hi
we have 2 solaris 10 servers in veritas cluster.
also we have oracle cluster on the database end.
now we have a requirement to reboot both the servers as it has been running for more than a year.
can any one tell what is the procedure to bring down the cluster services in both the nodes... (7 Replies)
Discussion started by: newtoaixos
7 Replies
7. AIX
Hi,
A customer I'm supporting once upon a time broke their 2 cluster node database servers so they could use the 2nd standby node for something else. Now sometime later they want to bring the 2nd node back into the cluster for resilance. Problem is there are now 3 VG's that have been set-up... (1 Reply)
Discussion started by: elcounto
1 Replies
8. UNIX for Advanced & Expert Users
Is there any way to create a arbitrary node for ocfs2 on a virtual machine (others are physical servers) so it won't go panic when one of physical server goes down?
This is for load balanced application servers.
Any setting example or tips?
Thanks. (0 Replies)
Discussion started by: malayo
0 Replies
9. Red Hat
Hi All;
I try to build a Redhat Cluster (CentOS 6) on vmware. But each node sees the other down like:
# clustat
Cluster Status for mycluster @ Wed Apr 8 11:01:38 2015
Member Status: Quorate
Member Name ID Status
------ ---- ... (0 Replies)
Discussion started by: Meacham12
0 Replies
10. Red Hat
Hi All;
I try to build a Redhat Cluster (CentOS 6) on vmware. But each node sees the other down like:
# clustat
Cluster Status for mycluster @ Wed Apr 8 11:01:38 2015
Member Status: Quorate
Member Name ID Status
------ ---- ... (1 Reply)
Discussion started by: Meacham12
1 Replies
LEARN ABOUT HPUX
pagezero_daemon_enabled
pagezero_daemon_enabled(5) File Formats Manual pagezero_daemon_enabled(5)
NAME
pagezero_daemon_enabled - zeroing of free memory in the background is enabled
VALUES
Default
Allowed values
Minimum:
Maximum:
DESCRIPTION
HP-UX provides improved security by zeroing out any memory that is being assigned to user space. This ensures that no user can read what
may have been written by some other user. Normally, the zeroing is done when the physical page is allocated to the user -- often when the
application touches the page for the first time. Some system calls like also result in the zeroing of pages. The time taken for these
kind of system calls and accesses depends upon the size of the memory being zeroed out. A 4G page may easily take many seconds to be allo-
cated. A large database shared memory segment may take many minutes to be allocated. Allocation of a small page is generally unobserv-
able.
The daemon is a performance enhancement that reduces the elapsed time for completing kernel operations like page faults, and so on. The
idea is to zero out large free pages (4MB and above in size) during times when the CPU is idle. The daemon has been specially designed to
ensure that it only executes for brief times when the CPU is idle.
However, under certain conditions, where some resource (CPU, TLB or Memory bandwidth) is very highly utilized, it is possible that the
operation of the daemon adversely impacts performance. For example, if the application is limited by the memory bandwidth, then it may be
better to disable the daemon. This situation is expected to be rare. Most workloads will not need to disable the daemon.
This tunable allows system administrators to disable and enable the daemon. Once the daemon has been disabled, it will not zero out any
more pages. Any pages already in the process of being zeroed will be zeroed out. When the daemon is enabled, it will zero out any unze-
roed free pages of size 4MB and more.
Who is Expected to Change This Tunable?
Anyone.
Restrictions on Changing
Changes to this tunable take effect immediately.
When Should the Value of This Tunable Be Changed to 0?
This tunable should be changed only if the system is using large pages (i.e. is set to 4MB or above). Changing it if the system is not
using pages of size 4M and above will have no effect. It may be changed to 0 if the system is bottle-necked on some hardware resource. In
particular, disabling the daemon may help in decreasing TLB misses and increasing the available CPU time and reducing memory latencies.
What Are the Side Effects of changing the value to 0?
Disabling the daemon will increase the time it takes to process page faults and to complete system calls that result in the allocation of
memory (e.g.
WARNINGS
All HP-UX kernel tunable parameters are release specific. This parameter may be removed or have its meaning changed in future releases of
HP-UX.
Installation of optional kernel software, from HP or other vendors, may cause changes to tunable parameter values. After installation,
some tunable parameters may no longer be at the default or recommended values. For information about the effects of installation on tun-
able values, consult the documentation for the kernel software being installed. For information about optional kernel software that was
factory installed on your system, see at
AUTHOR
was developed by HP.
SEE ALSO
vps_ceiling(5).
Tunable Kernel Parameters pagezero_daemon_enabled(5)