UCS Server CPU Allocation

Dear Guys...
As i am aware aabout UCS server and UCS system..
one physicla cpu is equal to virtual cpu...
but as i seen many application that having more that 6 cpu need i want to run on one UCS sever...how can adjust the cpu for them?
As i have given the BOQ below, it is having 2 servers and 4 CPU but i need to isntall many application:
CUCM
CCUCN
CCX
WEBEX
presence
SO it takes around 5-6 cpu for installation...and also 1 cpu per vmware software also need...
so that can you help me to correct for this issue and easy way to understand and configure the correct ucs server for my applications...?
Thanks and Regard's
Amit Sharma

Amit,
Please start from here for running UC apps on UCS
http://docwiki.cisco.com/wiki/UC_Virtualization_Supported_Hardware
http://docwiki.cisco.com/wiki/Unified_Communications_VMware_Requirements
HTH
Padma

Similar Messages

Mixing BE6000 UCS Server and "normal" UCS server in the same deployment

Hello,
I have been handed a project which has one high density BE6000 UCS server and a separate UCS C220 M3 server. The latter server was included to host a MediaSense call recording system but this will only use 2 of the available 8 vCPUs on the UCS C220 M3 server.
The total number of users is 200.
I want to implement a resilient system and so would like deploy two servers in a cluster for each of the following applications:
CUCM 10.5
Unity Connection 10.5
IM & Presence 10.5
As well as these applications there will be UCCX 10.6 and Cisco Unified Attendant Console Advanced (10.5) but these will be deployed as single servers.
Looking at the UCS servers they have capacity for me to split the CUCM/CUC/IMP clusters between them.
I cannot see any technical reason why this will not work but do not want to be caught by any Cisco support policies.
If I were to implement the system in this way would there be any issues with the deployment or getting support from TAC.
The separate UCS C220 M3 server has 8 x 8GB RAM (64 total) and 8 x 300GB HDD plus a quad Ethernet card.

James first of all I am not sure I understand your query in detail. Do you mean you have two UCS220 M3 servers? and one of them is currently running BE6000?
Having said that, the key here is to carefully plan your deployment against the capacity of the servers you have,.
Eg..Deploying 200 Users, using the ff OVA
UCS 220 M3 server 1: (using default TRC ie 8vCPU with 8GB per vCPU)
Publisher (2500 cucm OVA): 1vCPU (6GB RAM), 80GB HD
IMP-Publisher (1,000 OVA) :1vCPU(2GB RAM), 80GB HD
CUC-publisher (1,000 users OVA) :2vCPU(6GB)--NB 1vCPU is reserved for ESXI, 160GB HD
UCCX-Master: (300 agent OVA): 2 vCPU (8GB) and 292GB HDD
With this placement you have a total of 22GB RAM and 612HDD used up
A break down is shown below..
Server
Server Name
C220 M3S TRC#1 (Medium)
C220 M3S TRC#1 (Medium)
Application Short Name
Application Long Name
Release
VM Name
vCPU
vRAM
vDisk
CUCM
Unified Communications Manager Release
10.x
CallCtrl: 2,500 users
1
4
80
IM&P
IM & Presence Release
10.x
1,000 users
1
2
80
CUC
Unity Connection Release
10.0
1,000 users
1
4
160
ESXi
Unity Connection
ESXi
1*
CUCCX
Cisco Unified Contact Center Express / Unified IP IVR Release
10.x
Main: 300 agents
2
8
292
ESXi
VMware vSphere ESXi
5.5
4**
* Note: This is a 1 physical CPU core per host regardless of the number of Cisco Unity Connection (CUC) VMs.
** Note: This is 4GB physical RAM per host.

Error message of MAC flapping for CCM-UC installed in UCS server

Dear All,
I have a strange issue, I have the 2 UCS servers setup with CCM & unity Connection running on each server on PUB - SUB mode.
I'm getting following error message on the Switch where the UCS server Connects.
I'm unable to trace why we are getting this error message, There is no problem with accessing CCM or UC, also the Phone & Voice mail seem to working.
un 25 13:15:14.853: %SW_MATM-4-MACFLAP_NOTIF: Host 000c.2911.db00 in vlan 410 is flapping between port Gi1/0/34 and port Gi1/0/33
Jun 25 13:15:28.828: %SW_MATM-4-MACFLAP_NOTIF: Host 000c.2911.db00 in vlan 410 is flapping between port Gi1/0/34 and port Gi1/0/33
Jun 25 13:15:44.145: %SW_MATM-4-MACFLAP_NOTIF: Host 000c.2911.db00 in vlan 410 is flapping between port Gi1/0/34 and port Gi1/0/33
Has anybody expereinced such error before, Appreciate your help in tracing this error.
Thanks,
Oliver

Hello,
I believe you'll get better answer at:
"https://supportforums.cisco.com/search.jspa?peopleEnabled=true&userID=&containerType=14&container=2003&spotlight=true&q=UCS"
my best guess is to check the LAN config on the UCS. It seems its trying to load balance the traffic in Active-Active rather than Active-Standby.
HTH

How many cpus on server are allocated to client

hello,
its not an issue rather doubt which is hunting me from last 2-3 days.
My 10gR2 Oracle database running on
Widows server 2008 R2 Enterprise
Processor ----> intel Xeon cpu [email protected] GHz (4 processor)
RAM------>8GB
its a dedicated server.
Actually i was going through the AWR report whn one of the client came to me with slow database performance during some time interval.
when i had a look at the Top 5 Timed Events in Report it showed me the
Sql*net message from client      834.35si told them to run their query so that i can have trace of that.
then time event showed me something nearer to above and our database was taking only 15 sec to carry out the task.
i told them that
"you application/client is taking 834 sec to tell our database to do something.But once the database got your application's request to carry out something, the database is taking only 15 sec to provide you with the result"
so they asked me to justify that the problem is with application side/machine side.
After digging in a lot on the internet ground, chatting with my colleges on cell phone and running through many threads specially Asktom.oracle.com, i got really confused.
They came up with varied opinions which kept me thinking again & again over the same thing.
In one of his(Thomas kyte) thread the discussion was as follows.
The output of tkprof when the developers run the same program on their PC shows a big difference in
SQL*Net message from client.
The SQL*Net message from client from my tkprof is almost 4-5 times theirs.
Could you please advise what might be the cause of that?
Thanks
Followup   August 25, 2009 - 9am Central time zone:
that could mean.....
your machine is slow (their desktop cpus blow your server cpus away) as much as anything.
since their code is not instrumented, try SQL*Net trace (with TIMESTAMPS) can be helpful to check out the time in the network between the client and server. See Note 16658.1 - see if you experience radically different response times on the various machines.
4 stars   August 25, 2009 - 2pm Central time zone
Bookmark | Bottom | Top
Reviewer: Jaklin Ekdawi
My Database server is T5120 with 4 core 1.2 GHz Ultra SPARC T2 processor (from O/S it is seen as 32
CPU * 1.2 GHz).
Their PC has 2 CPU * 4 Ghz.
Based on the above, How come their PC is more powerful than my server?
The ASH Report for the execution period of the program shows ¿CPU + Wait for CPU¿ is the top event,
however, tkprof shows ¿SQL*Net from client¿ is the top event waited on.
Top User Events
Event        Event Class    % Activity    Avg Active Sessions
CPU + Wait for CPU     CPU             81.66        0.14
log file sync             Commit             5.92        0.01
db file scattered read     User I/O     2.37        0.00
Can you please explain why there is a difference between tkprof and ASH report?
Thank You
Followup   August 25, 2009 - 8pm Central time zone:
I only care about a single cpu, your client program is not using more than one.
I often downloaded stuff from my big bad sparc to my desktop because my desktop was many times faster than the sparc.
look further down in the ash report - sqlnet message from client is an idle wait, we tend to ignore it as it is the time we spent WAITING FOR THE CLIENT.
Their pc looks a lot faster than your server since you are only using a single cpu - your client is not multi-process. It only uses a single cpu.
4 stars   August 25, 2009 - 2pm Central time zone
Bookmark | Bottom | Top
Reviewer: Jaklin Ekdawi
Sorry, I forgot to say that during the execution time the cpu on the server was 98-99% idle based
on (using top and sar).
Thanks
Followup   August 25, 2009 - 8pm Central time zone:
I agree, because you could not use 31 out of 32 cpus, I would expect it to be almost idle.
97% idle would mean a single cpu was 100% used. So, it sounds like you were using almost 100% of that single cpu you had access to.
and it isn't as fast as their pc.
4 stars   August 25, 2009 - 10pm Central time zone
Bookmark | Bottom | Top
Reviewer: Jaklin Ekdawi
The tkprof from the developer's PC is below
OVERALL TOTALS FOR ALL NON-RECURSIVE STATEMENTS
call     count       cpu    elapsed       disk      query    current        rows
Parse      122      0.39       0.68          0        403          0           0
Execute    122      0.00       0.00          0          0          0           0
Fetch      396      0.09       0.11          0      25354          0         274
total      640      0.48       0.80          0      25757          0         274
Misses in library cache during parse: 122
Elapsed times include waiting on following events:
Event waited on                             Times   Max. Wait Total Waited
----------------------------------------   Waited ---------- ------------
SQL*Net message to client                     763        0.00          0.00
SQL*Net message from client                   763       14.71        312.91
SQL*Net more data to client                   122        0.00          0.00
The tkprof for the running the same process on Solaris DB Server
OVERALL TOTALS FOR ALL NON-RECURSIVE STATEMENTS
call     count       cpu    elapsed       disk      query    current        rows
Parse      124      0.06       0.03          0          0          0           0
Execute    124      0.04       0.01          0          0          0           0
Fetch      400      0.07       0.10          0        802          0         276
total      648      0.17       0.15          0        802          0         276
Misses in library cache during parse: 1
Elapsed times include waiting on following events:
Event waited on                             Times   Max. Wait Total Waited
----------------------------------------   Waited ---------- ------------
SQL*Net message to client                     774        0.00          0.00
SQL*Net message from client                   774       43.65        964.23
- On Solaris DB server, the CPU is 0.17, however, "SQL*Net message from client" is 964.23.
- On the developer¿s PC, the CPU is 0.48, however, "SQL*Net message from client" is 312.91.
Based on that, how I can approve it is CPU issue?
Followup   August 26, 2009 - 7pm Central time zone:
how about this.
have the developers write something that mimics what they do without the database in it. Same sorts of loops and processing they do in the code in the client.
run it on their machine
run it on yours
You'll find your big bad server has cpu's that are extremely slow compared to the desktop machine.
Hence - the time spent waiting for sqlnet message from client is - time spent in YOUR client on YOUR server processing the data returned from the database.
saying "the cpu 0.17 on one and 0.48 on the other" is meaningless, they are completely different cpus with completely different performance characteristics.
And that you have 32 of them - no use to you, you use one..
It looks like your server is about 2-3 times slower then their test machine.
5 stars SQL*Net Message from client may not be an issue from the app   May 28, 2010 - 8am Central time zone
Bookmark | Bottom | Top
Reviewer: Thierry B from Paris, France
Hi,
I just read your post as I experienced the exact same problem. I found something that fixed this
problem for me, my sqlnet.ora contained that line after an oracle upgrade:
TRACE_LEVEL_CLIENT=16
I don't know how it came there. But it was the cause of my problem (too long idle time). I simply
removed it and no more wait events.
I hope it can be of any help for anyone.
Followup   May 28, 2010 - 8am Central time zone:
that is a client side issue though, isn't it. It is an application side issue, not a database side issue.
5 stars   April 15, 2011 - 10am Central time zone
Bookmark | Bottom | Top
Reviewer: A reader in this thread Tom insisted that the client has access to only 1 CPU on the server even though the questioner told him that his machine is having around 32 CPU'S.
so this thread was the source of my confusion.
i kept asking myself "if the client application has access to one cpu out of 32 cpu's on the server on which the database server is running,then What is the use of multiprocessor? my oracle/client application has access to only one cpu on the server(like Tom said), what other 31 cpu's on the server machine are upto?if one cpu is getting exhausted by the client Application request, then what the other processor are doing ? are they not meant for load balancing?"
when i had a chat on cell phone with my colleagues regarding the oracle access to the cpu's on the server, then they told me that if there are more number of cpu's then the the response time would be fast cause workload is divided among the multiple processor to carry out the task in parallel.
Again i got confused with the deferring opinion..
so just want to have clarification on
* how many processes do i have access to on server machine where my oracle is installed?*
how many processes do the client applicaiton S/W/client machine have access to out of 32 processors?
is workload devided among the cpus on the server ,in case of huge request from the client application?
thanks & i hope my doubt would be cleared

Look at it in basic terms (forgetting for the moment about Oracle and SQL*Net and specific s/w itself).
A process is loaded into memory. This process contains machine code instructions. How many CPUs can execute these instructions?
Only 1.
Simplistically. The process has a single execution pointer and set of CPU registers. A single CPU executes the instruction (as indicated by the execution pointer). The registers are used and updated.
The same CPU may not be used each time around. So a single process can during its lifetime, be executed by a number of CPUs. But a single CPU at a time.
A process can thread. This basically means that a second copy of the execution pointer and registers set are created. Each such copy will be executed by by a single CPU at a time.
An Oracle client is serviced by an Oracle server process (either dedicated or shared). On Linux/Unix systems, this will be a unique physical process. On Windows, it will be a unique thread.
This process is executed by a single CPU. Thus that client is serviced by a single CPU.
This changes when Oracle can use parallel processing and there are PX slave processes available. In that case, the client is serviced by a single process - but that process has in turn enlisted the assistance of helper processes. Thus the client is now serviced by multiple processes (or threads on Windows) and thus by multiple CPUs.
However, it is not that simple either. It is "cheaper" for the kernel to run threads on the same CPU than on different CPUs. The reason is the same memory is used by both and access to that memory by different CPUs is more complex (and thus more expensive and slower) than from the same CPU.
So depending on a number of factors, that client may still be serviced by a single CPU despite the fact that PX slaves are used. And from a client process perspective, this does not really matter. The crux of the issue in this regard is the CPU utilisation/footprint of that client process on the server.
If the clock time for a client process says "+running for 60s+" and the server says "+CPU time is 10s+", you cannot say that the process was not serviced by multiple CPUs. You do not know whether than 10s CPU time was for 5 processes each spending 2s to service that client.
However, when the server says "+CPU time is 70s+" - a longer period that the wall clock time of the client, then it is reasonable to assume that multiple server processes (and thus CPUs) are servicing that client process.
As for exactly how many CPUs are in fact used by the server? That has no real bearing on this. It does not change client-server behaviour.
Even Oracle does not know and does not care which CPU is servicing it. The correct piece of s/w to deal with that, is the kernel. Yes, you can bind a thread or process to a specific server CPU, but the reasons for this is quite technical and is always an exception. Let the kernel deal with which CPU to use. It is of no concern to the server s/w (like Oracle), and even less to the client process being serviced by that server s/w.

Server CPU Usages - Critical

We have a BC4J application developed using jdev 3.1/ front end is JSPs.
The application is deployed using Oracle iAS 8i.
The problem is that when in a jsp, a query is executed that has to return a large number of rows, then the cpu utilization of the database(oracle.exe) rises to 100 percent. The same query when run from sql, causes a memory consumption of only 20 percent.
What can be the possible reasons for this huge increase in memory consumption when the query is fired using jdev data web beans from a jsp.
amit

Howard,
because the CSS has a script language, you can basically do whatever you want.
You could have a script running on the server to get the server cpu and save it into a file.
Via the CSS script language, you could retrieve this value and decided to use or not the server.
There is also something called DFP - dynamic feedback protocol. If your server vendor support this protocol you can use it to report server status to the CSS.
Regards,
Gilles.

Oracle 11g - Solaris 10 on cisco UCS server

I would like to know if anyone has experience with installing Oracle databases on cisco UCS servers. I recently took over a dba shop and my client has purchased a cisco UCS server and is planning on migrating some databases currently on dedicated servers running on Solaris 10 and others running on Linux RH 5.4 platforms. I need to find out if Oracle 11g and Solaris 10 is compatible and covered by Oracle licenses. Does anyone have any specs and/or information on this topic? Thanks in advance.
Jonathan Begazo-Leon

Jonathan,
there's a dedicated Oracle database forum (Database) where you can post your issue. In this forum here only 3rd party to Oracle migrations using the Oracle tools migration workbench or SQL Developer will be covered. As I can't move your thread, please close this one here and post your issue again in the database forum.
Thanks,
Klaus

RAID 1 is not showing in UCS Server configuration utility

Hi,
I have a UCS Server C200 M2 Rack Mount and planning to install Windows 2008 R2 server in it. To make it High Available, i m configuring RAID 1 in it.
I can easily configure RAID from BIOS setup and after rebooting it shows me that the RAID 1 is found and Online in status.
Once I insert the Cisco UCS Server Configuration Utility 2.0 CD and try to check the RAID Configuration, it show me error message "SCU is unable to detect a RAID controller in the server"
If I move further to install OS then during installation setup gets hang and system reboots.
I attached the screen shot for review and help.
Kindly advise!
Thanks,
Aman

Hi Aman,
Can you download the Driver package from
http://www.cisco.com/cisco/software/release.html?mdfid=283860950&flowid=25801&softwareid=283291009&os=Windows%202008%20R2%20(x64)&release=1.3(1.1)&relind=AVAILABLE&rellifecycle=&reltype=latest
And extract w2k8r2_intich10r_13.21.0614.2010.zip for the Intel ICH10R integrated RAID controller.
Try Browsing to this Driver either from the SCU or during the direct OS reinstall and make sure
"Hide drivers that are not compatible with hardware on this computer" is unchecked to be able to point to the Driver
I assume you are seeing the problem as described in the RAID configuration issues in
http://www.cisco.com/en/US/docs/unified_computing/ucs/c/ts/guide/TS_2Server.html
Installing Windows 2008 64-bit and RAID controller had issues.
LSI drivers are not bundled in Windows 2008 64-bit. These must be installed during the installation process. During the install process, if the hard drives fail to be detected, use the load driver option to point the drives to the correct drivers for the LSI controller in the system. The drivers can be loaded using a USB drive. When loaded, the hard drives are displayed and the hard drive for the OS can be selected.
Please let me know how the installation goes.
Regards,
Gaurav Samanta

Server CPU as Keepalive Parameter - Is it possible?

Is there anyway for Keepalive to check the server CPU utilization as a way to determine if the service is dead or alive?
Regards,
Howard

Howard,
because the CSS has a script language, you can basically do whatever you want.
You could have a script running on the server to get the server cpu and save it into a file.
Via the CSS script language, you could retrieve this value and decided to use or not the server.
There is also something called DFP - dynamic feedback protocol. If your server vendor support this protocol you can use it to report server status to the CSS.
Regards,
Gilles.

Can we install PGW 2200 Softswitch on UCS Server?

Dear friend.
I intend to build a backup system for my customer using PGW2200 Softswitch.
I have read that it should be install on Sun Netra X4270 platform. So can I install it on UCS server? Have anybody have experence in this system.
Thanks for your time
Vinh Phan

On UCS server it supported 9.1 and after named as vPGW. So you have to install Solaris OS first and top it vPGW application .
Any degraded version would not supported on UCS server.
Regards
Nadeem
Sent from Cisco Technical Support iPhone App

CUCM - UCS Server Replacement

Hi All,
We have installed CUCM 9.1.2 on UCS 220-M3 server (standby hardware). Now we are going to replace the UCS server.
Has any other vmware method or tool available to migrate the virtual machines one UCS to another UCS ? { VM's are resided in DAS) ?

Technically, Cisco only supports it if using shared storage- http://docwiki.cisco.com/wiki/Unified_Communications_VMware_Requirements#VMware_vMotion
But it usually works fine from a technical standpoint. Just might not get support if something goes wrong with the migration.

Monitoring on UCS server (ESXi)

Hello all,
I am wondering normally how we monitor UCS server( I mean ESXi).
The virual machine could be monitoring via SNMP.
vCenter could provide us system status.
Could we set up SNMP on VMWare ESXi ? So we could monitor through HP Openview etc.
Thanks a lot,
Howard

Hi,
I'm not sure if HP openview is supported, but you can refer to the UCS SNMP and MIB configuration guides below:
http://www.cisco.com/en/US/products/ps10281/products_configuration_example09186a0080b36928.shtml
http://www.cisco.com/en/US/docs/unified_computing/ucs/sw/mib/reference/UCS_MIBRef.html
Regards,
Tere.
If you find this post helpful, please rate!

What is the procedure of Call Manager upgrade from version CUCM 7.1 (MCS Server) to version CUCM 10.2 (on UCS server)

Hi
What is the procedure of Call Manager upgrade from version CUCM 7.1 (MCS Server) to version CUCM 10.2 (on UCS server) , please let me know the steps to what would be the backup procedures and upgrades.
Regards
Gaurav

I'd suggest reading this guide to make sure you have prepared your environment to support 10.x first. The link below also contains the upgrade path to 10.0(1).
One of the ways you can do this is:
1/ Apply refresh upgrade cop file on all servers in cluster.
2/ Upgrade the cluster to 8.6(2). Take a backup of the cluster.
3/ Build a cluster of 8.6(2) servers on VMs, and assign them the same hostnames and IP address as the hardware cluster. (You may need to keep these on a separate network until switching off the hardware machines)
4/ Restore the backup taken from the hardware cluster on to the VM cluster.
5/ Upgrade VM cluster to 10.0(1).
http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/upgrade/10_0_1/CUCM_BK_U4214F9D_00_upgrade-guide-cucm-100/CUCM_BK_U4214F9D_00_upgrade-guide-cucm-100_chapter_010.html#CUCM_RF_UA60AFAB_00

Application Server CPU at 100%

We are running Oracle Forms and Reports 9.0.4 using OC4J on a Solaris 9 server. Environment responds quickly, except when an end user requests the generation of a large PDF report (with several hundred to over a thousand pages). Although query results are returned quickly by the database server, CPU on application server tops out at 100% for an hour or more while these large reports are being generated. During this hour or more, no other reports can be generated by other end users. Can the application server be configured to allow CPU and other resources to be shared during generation of large PDF reports? If so, how?

>
... -Dlogfile=/vol04/applmgr/DEV/inst/apps/DEV_devas/logs/appl/conc/log/FNDOPP216337.txt ...
>
This looks to be the XML/BI Publisher process. Can you post relevant sections from this file ?
Pl post EBS related questions in the EBS forums - http://forums.oracle.com/forums/category.jspa?categoryID=3
HTH
Srini

Oracle export process eating up 95-99% of server cpu

Hi Everyone,
I noticed that my export process is eating a huge amount of the server's cpu resources.
Is there a way to control the resources allocated to my export process?
thanks

>Isn't that spinning and CPU usage ?
Yes. No. Spinning on a latch means that the process is in wait state - waiting to access the resource. It means that every now and then, the process will check the resource state to determine if it can obtain access to it.
Does "spinning" eat up CPU time? Typically not. And 99% of the time where is does, it is due to the type of shoddy "loop spinning" that I've demonstrated with the 1st sample piece of code above (called [url http://en.wikipedia.org/wiki/Busy_waiting]busy waiting).
The norm is that a process in a wait state consumes very little CPU time. There are actual CPU instructions that tells the CPU to "ease off", like the [url http://en.wikipedia.org/wiki/HLT]HLT instruction. This instruction is used at a higher level abstract(API) level - like with the DBMS_LOCK.SLEEP() API call.
> CPU usage DOES convey information.
You are missing the basic point that I was trying to make. In terms of performance measurement, it says nothing. It does not say GOOD. It does not say BAD.
Why?
An idle CPU can be a result of exceedingly poor performance caused by process serialisation (like in PL/SQL doing your own sequences by locking a table to do a SELECT MAX() to get next sequence number to use). Seen this one myself and most recently on a Superdome that showed idle CPU utilisation with the performance being excruciatingly slow.
An idle CPU can be simply because of no work to be done.
An idle CPU could be because of an I/O bottleneck as a 200 parallel query processes all hit the same I/O channels.. and then have to sit back and wait for the I/O calls to be completed. Saw this one often on Oracle Parallel Server (OPS) many years ago.
Similar situation for a heavily utilised CPU. The actual utilisation figure does not tell you how much of that CPU time used, is used effectively by optimally designed and coded processes.. and how must is wasted by poorly design and coded processes.
From a performance perspective, it alone says nothing.
And this is what the OP says - he's seeing 95+% CPU utilisation when running an export. Is it a performance problem? If so, what can he do about it?
The answer to which is, identify if there is an actual performance problem as 95% CPU utilisation alone does not tell us whether that is good and efficient usage of CPU time, or abuse and wasted use of CPU time.

SQL Server NTFS allocation unit size for SSD disk

Hi,
I have read that the recommended NTFS allocation unit size for SQL Server generally is 64 kb since data pages are 8 kb and SQL Server usually reads pages in extents which are 8 pages = 64 kb.
Wanted to check if this is true also for SSD disks or if it only applies to spinning disks?
Also would it make more sense to use an 8 kb size if wanting to optimize the writes rather than reads?
Please provide some additional info or reference instead of just a yes or no :)
Thanks!

Ok thanks for clarifying that.
I did a test using SQLIO comparing 4kb with 64kb when using 8kb vs 64kb writes/reads.
In my scenario it seems it doesnt matter if using 4kb or 64kb.
Here are my results expressed as how much higher the values were for 64kb vs 4kb.
Access type
IOps
MB/sec
Min latency (ms)
Avg latency (ms)
Max latency (ms)
8kb random write
-2,61%
-2,46%
0,00%
0,00%
60,00%
64kb random write
-2,52%
-2,49%
0,00%
0,00%
-2,94%
8kb random read
0,30%
0,67%
0,00%
0,00%
-57,14%
64kb random read
0,06%
0,23%
0,00%
0,00%
44,00%
8kb sequential write
-0,15%
-0,36%
0,00%
0,00%
15,38%
64kb squential write
0,41%
0,57%
0,00%
0,00%
6,25%
8kb sequential read
0,17%
0,33%
0,00%
0,00%
0,00%
64kb squential read
0,26%
0,23%
0,00%
0,00%
-15,79%
For anyone interested this test was done on Intel S3700 200gb on PERC H310 controller and each test was run for 6 minutes.

UCS Server CPU Allocation

Similar Messages

Maybe you are looking for