Bottlenecks of Performance measurement & methods

Similar Questions

• You are looking for a performance measurement tool Labview?

  Hello

  In Labview, do we have a performance measurement tool or program that comes

  with Labview software?

  By the way, I have more LabView 8.6 development?

  Thank you

  Do

  Thanks to you all.

  That's exactly what I was looking for!

• VMWare performance - measurement (more than processor cycles)

  Performance of VMWare-

  Cluster:

  Guests: 2 proc, 4 core, Zeon 2.659 X 5550, 96 GB Ram: 5 Cluster Host (Total 106360 Mhz)

  Our cluster VM rest (45 to 65%) about 50 days/nights/weekend - why?

  No application server except -

  (2) Server MSSQL (10 access, low-bandwidth users)

  (6) the network of monitoring/taps-

  (3) ads - for 100 users servers

  (2) print for 100 users servers-

  (20) light Win - 7 offices (used for outlook only)

  (20) desktop Linux computers

  (8) VM used by Avamar backups/strip out

  (40) Server 2008/12 - do nothing if you check taskman local/performance monitor

  vCPU/Core = 6.89 to 1 = is too high?

  MEM: 455GB granted, 347GB consumed, 46GB active

  Why (how save/measure) charge a customer on a host? What is running in the client requesting the resource? that is how can I tell that the cycles used on my guests were actually used/required by the customer?

  How/where the disk/CPU/network are followed for backup using VADP API calls?

  Efficiency/effectiveness of the Hyper-threading? How can you tell if the customers take advantage of HT?

  You are looking for the best reason why our environment virtual machine uses 'X' amount of resources. If the answer is: "clients use resources they need" - that's fine, but the documentation would be nice.

  Thanks, Cliff

  vROPs is a unique virtual appliance.   It is a simple installation:

  • Deploy the FVO
  • Connect via the browser
  • Simple method to choose and choose the trial license
  • Log in as Admin
  • Return to Administration-> Solutions-> VMware vSphere-> vCenter adapter
  • Configuration of adapter vCenter (vCenter FQDN and credentials for vROPs to extract data from vROPs)

  She's.  Within 5 minutes, it will start by pulling all the metric vCenter.

  Data will be useful once there to meet your needs of analysis.  vROPs makes the dynamic thresholds, which take longer to begin generating, but you can watch the metric in the troubleshooting ASAP.

  -MattG

• Bottleneck of performance when importing datapump

  Hi, we have created a new database 12cR1 on LMOS with Linux x 86 64-bit. We are experiencing great slowness in 10-11GB data import and taking more than 3 hours. Looking at the AWR it shows high I/O waits.

  AWR attached but unable to conclude the problem. Please can I ask someone to stem the problem?

  Thank you very much!

  I haven't looked too closely, but I will note this:

  Your sql albums and waiting is all related to import - which is not a bad thing.

  You start on the cpu - that is not a bad thing, even if CPU resources are needed by the operating system to run scripts, so there could be something on the side of the BONE there to observe.

  You have a lot of 'other' waiting for I/O - I'll take that as meaning the OS did a lot of things.  It makes sense for a big load.

  Others may have more insight of the AWR in what Oracle can be done, but I don't think that we are not seeing clear is the resources required by the operating system to archive the redo.  Therefore, if no one else using the db when you perform this type of load, you might consider: backup, off the record of archive, load, turn the back of archiving, archive again.  You can use OS tools to check the performance in your current setup and configuration noarchivelog, and total elapsed time.  Given that the percentage of your database that processes these data compared to the full size will determine how long does take a full backup, this could be the determining factor to see if you do not want to do that.

• Performance measures

  Hello

  have a question - because it is confusing me in the moment.

  i ' AM trying calc "CPU Ready" of the summation of value to a more reliable value once daily, weekly, monthly analysis.

  From some I did some views overall - but output is not really to information of gui.

  If I analyze for example only 'real time' - I tried with 2 orders.

  (get-stat-realtime-stat cpu.ready.summation - entity xxxxxx | measure-object of value - average) subject

  and

  (get - vm xxxxxx | get-stat - cpu.dready.summation stat - realtime | measure-object of value - average) subject

  Then, I get:

  Average = 47

  If I compare real-time Gui > can I get (117).

  So do not understand the difference

  If I do the same query with "Maximum" I return the same value as in Gui

  Theire is a kind of mistake I did question the value average real time?

  THX

  Max

  
  

  Run the following

  Get-VM xxxxxx | Get-Stat -Stat cpu.ready.summation -Realtime |Group-Object -Property Instance
  

  You'll see at least 2 groups in the output.

  According to the number of vCpu you configured for this virtual machine, you will see more groups.

  You have the group with the value "in white", which is the aggregate value (for all the vCPU in the virtual machine).

  And then you a group with a CPU id as the value, it is for each vCPU virtual machine.

  You want the overall value (this is the one you are most likely see in the GUI).

  Get-VM xxxxxxx | Get-Stat -Stat cpu.ready.summation -Realtime |Where {$_.Instance -eq ""} |Measure-Object -Property Value -Average | Select -ExpandProperty Average
  

• How performance measurement/reference of a new database on the new server?

  Hello

  I have two servers oracle with following (same) details:

  5.8 64-bit RHEL

  Oracle 10 g 2-10.2.0.5.8

  ASM 10 gr 2-10.2.0.5.8

  Server A: 32 GB of RAM, 8 processors @ 3 .00GHz

  Server b: 128 GB RAM, vCPUs 16 cores

  A server (physical server) already has a database of A. ServerB (VMware - Yes, my client moves all Oracle on VMware servers for some reason any) is a new installation and the new database B with exact same init params as A databas. I expdp data from database A and impdp in database B.

  According to the material, the material is better than the old server B. I made a very simple test to check if the new DB performs better than the one on the physical server. Here are the results:

  I ran a simple query to create a new table. The original table (say, table_a) contains 1.7 million lines and the size is 2.2 GB.

  create table test1

  as

  s elect * from table_a;

  It took 3:28mins on database B, it took 1:55mins only on A database. So, it seems that the new database B scene poor (apparently). Then I looked at the plan of the explain command (don't know if it means a lot, because it s a very simple query) and here's these two databases:

  A (physical server databases

  Plan
  SELECT STATEMENT ALL_ROWS Cost: 14 052 bytes: 2,161,302,003 cardinality: 16,250,391
  	1 FULL TABLE ACCESS TABLE table_a Cost: 14 052 bytes: 2,161,302,003 cardinality: 16,250,391

  Database B (virtual server)

  Plan
  SELECT STATEMENT ALL_ROWS Cost: 59 844 bytes: 2,161,302,003 cardinality: 16,250,391
  	1 FULL TABLE ACCESS TABLE table_a

  
  Cost: 59 844 bytes: 2,161,302,003 cardinality: 16,250,391

  Issues related to the:

  1. Why is the cost different? Should I "calculate statistics" on the B (virtual server) database?

  2. how to pursue the investigation and discover the reason for the difference in time?

  3. what other test can I run to make sure that I have the right database Setup?

  Don't know if it of just enough info - if not, please let me know what else I need to provide.

  The team I perform hands on this server refuses to accept it by saying that it is slower than the existing one.

  Please advise!

  Best regards

  
  

  Looks like you'd get only a few percentages of increase in performance by increasing the pga, while the downturn is probably elsewhere.  Follow Nikolay, Ed and John suggestions.  STATSPACK or AWR can also give clues to e/s, although it would be a mistake to expect them to reflect the performance of the production - but they might highlight if you have problems logwriter or something.    Don't know if I even have to mention My ISP4400: numbers of ORION

• ESX/VM performance measures

  Hello

  I'm looking for a document that will help me establish metric values of acceptable performance for my ESX hosts and VMs. This would be the CPU, memory, disk, and network.  For an example and acceptable time ready for a computer processor virtual would be 1000 ms.

  Would you know if I can find something for other performance metric parameters such as disk latency, swap swap rate/rate, etc?

  Any help would be appreciated.

  Concerning

  rm0107

  See: http://www.yellow-bricks.com/esxtop/

  André

• XW 6400 processor upgrade

  Hello world

  I'm looking to upgrade the processor in my HP XW6400 workstation. I currently have 2 dual-core Intel Xeon 5130. Should I upgrade to the processor dual-core 5160 (3 GHz) or the quad core 5320 (1.86 ghz)?

  I think (if I understand correctly starting from a previous thread) that my motherboard supports up to the intel xeon 5320 and not the 5345 because it has a SP:436925 - 001 motherboard.

  Any thoughts would be appreciated.

  Thank you very much

  Aaron

  Aaron,

  It took some time to gather my data bits.  Here's the info: with Windows 7 Microsoft came out with a standardized performance measurement method can be run and see how your configuration batteries against others.  It comes to memory, CPU, graphics card and game performance and hard disk performance.  The test is called experience Windows Experience Index or user Windows and top scores on to 9.0.  You can get to right click on my computer in W7 and drag it down to the properties.  I have attached the results for the xw6600 I'm working on right now that has an older Intel SSD for its primary drive.  If you have two video cards the results reflect only one slot PCIe x 16 superior.

  It turns out that your hard drive partition can never exceed 5.9 without an SSD.  But what you really want right now is the scores of processor and memory, as the best indicator of the user experience practice.  For cards mothers xw6400 scores memory and processor are always to be balanced and identical, in all my tests.  Not so much for other HP motherboards, but the 6400, 6600, 8400 and 8600, this is the case.  I have, as you can guess, I was able to run the same test on several different HP workstations.  You will be able to extrapolate from the numbers below and will be amazed how the xw6400 is with these faster quad cores, I recommend.  Here are the data, weaker performance on site a group that I have personally tested.  The 5100 series processors are all two hearts; 5300 series is quad-cores.  The other odd are quadri-hearts of different families of Xeon in different workstations, as well as the a 6-core in a Z600 (which has a place for a second processor, unused, in this particular facility).  All the xw6400/6600/8400 and 8600 scores are made using the same label HP Server/Workstation memory, interchangeable, which is HP origin for each of these workstations:

  1 x E5140 = 5.8 processor and memory 5.8 scores, and I'm going to shorten things down to the only identical score from here on the results of CPU/memory in WUE.  It actually also got a 5.7 several times during the tests, so it's a 5.8 'low'... almost a 5.7.

  1 x E5160 = 6.4

  1 x = 7.1 E5345

  1 x E5430 = 7.2 (in a xw6600)

  2 x E5150 = a 7.2 'low '.

  2 x E5160 = a 'high' 7.2 that turned out as a 7.3 a couple of times

  1 x = 7.3 (in a xw6600) E5450

  1 x X 5355 = 7.3 (in a xw8400)

  1 x E5645 = 7.4 (in a Z600)... it is a single processor with 6 cores

  2 x 7.5 = E5345 (it's my favorite build if the xw6400 can run this image.)  15% cannot, in my experience.  I don't use any any Quad cores as this one.  I go to 2 x E5160 if I can't launch an E5345 quad core, and if I can run 1 of those so I was always able to run two.  SLAEJ sSpec code is the one to buy.  If you buy a processor used off eBay at least take the time to find the sSpec codes and get the last one for this processor).

  2 x X 5355 = 7.6 (in a xw8400)

  2 x = 7.7 E5450 (in a xw6600, that I'm working at this time)

  2 x E5440 = 7.7 (in a xw8600)

  None of the non-xw6400 processors listed above will run in a xw6400 with modern BIOS, said by the way.

  So, there is the data, and you can see that 2 x E5130 would be a dog compared to the use of 2 x E5160.  I understand that we get here cheaper access to of the E5345, and my buddies just come with their box of 6400 and me have plug in one of my E5345 checked to see if they should buy two of them, or two of the E5160.

  I have zero input for you on the slightest quad of the carrots in the list for the xw6400 HP... life is quite complicated, and I opted for this formula a / or for the 5160/5345.

  I hope this helps!

  Scott

• A measure of speed high speed with encoder in quadrature and NI 9401 on cDaq

  Greetings,

  We use an encoder in quadrature with 360 pulses/turn on the tracks (track A and B) and no trace of Z to measure motor speed at startup. Data acquisition, we use a NI 9401 in 9178 cDaq chassis and a pc with LabVIEW. The problem is that the start-up period is relatively short (less than 1 second), during which we measure speed as precisely as possible. The speed range is from 0 to 10000 RPM.

  What type of measurement method that you would recommend.

  Here are a few methods that we have already tried:

  -Measure with DAQmx CIFreq--> high frequency with 2 counters: speed measurement, but with a very big mistake (+ 166 RPM).

  -CIFreq DAQmx--> wide range with 2 counters: good speed data but more slow measurement,

  -CICntEdges DAQmx (counting separated the two lanes, speed conversion): very incoherent speed data.

  Thanks in advance for your help.

  Matej

  I would definitely say a 4, the measure of a low freq called option with 1 meter.  (Frankly, I've never been

  fond of this name because it is useful for freqs much higher than what I expect most people think "low freq".)   This

  is the method that I almost * always * use for frequency of counter measures.  It works really well to capture transitional

  variations in speed.

  10000 rpm and 360 cycles/rev, you are looking at a maximum frequency of 60 kHz.  The frequency measurement mode 1 meter

  There will be 80 MHz internal clock by encoder cycle edges, then you will get more than 1000 strokes per measure.  The point

  that means only 1 number of quantization errors, you can expect<>

  Further, you can average overall, say, 10 samples to you give even better accuracy and you could still be a data capture

  rate significantly higher than the probable bandwidth of your mechanical system.  (The average would just clean the jitter and noise and would not

  Hide answer true mechanical characteristics).

  -Kevin P

• Data Guard Performance: apply Lag

  In Cloud control there is a measure called "Lag (seconds) apply" under a group called 'Data Guard Performance '.  I created an alert on this metric rule.  However, I want to know what thresholds are defined for this measure, if I look under the supervision of models.  He isn't here.  There are also a number of other Performance measures of custody of data that is, I can't find or add in templates.  Can anyone help?

  From the home page of a database on hold, you can navigate to the (Oracle database > surveillance >) metrics and Collection settings page and select all the metrics view to check or set thresholds for (Data Guard Performance) metrics apply Lag. Note that by default, no limit is set for this measure.

  Kind regards

  -Loc

• Cannot display tables of performance of ESXi 4.1 data store

  Using Vcenter, I select the tab performance on each host ESXI4.1. I can view a chart for reading of data store and the latency of writing in real time. However, if I select any period other than in real time, the array is empty.

  Others for example CPU performance measures work well in real time and measures 24 hours a day.

  I have to be able to measure the reading of the data and store latency of writing on at least a 24-hour period. Someone had this problem?

  Examples shown in attachemnt.

  Hello

  Try to change the statistical level in 3:

  go to Administration - vCenter Server Settings - statistics and change the statistical level to 3

  wait for a while and check the historical performance, don't forget, change it back when you did with analysis

  See you soon
  Artur
  

  Please, do not forget the points of call of the "useful" or "correct" answers

• Call a method on InDesign full boot, how?

  Hello!

  I have not found anything in the documentation about this problem I want to solve. Basically, I need to call a method after InDesign has finished putting on and all user interface elements are ready.

  I looked in IStartupShutdownService, but this call my method too early when not all user interface elements have been upward and therefore results in a null pointer access. I'm sure I can't use CmdUtils::CreateCommand either since I only need InDesign to call a method and not change anything about the document or the template.

  A second option would be for me to know if my plugin Panel is visible after starting InDesign has completed and if it is displayed then perform this method call.

  I use myIdleTask:CIdleTask kSessionBoss AddIn.

  I can get the first call of RunTask() as a full boot.

  See sdksamples/paneltreeview/PnlTrvIdleTask.cpp.

• Sloow SAN performance. How can I confirm the vmfs format?

  Best practices says format SAN storage through the VI client.

  This is done for the blocks and the SAN pieces line up properly.

  It also provides more efficient storage for initial commissioning management and ongoing management of shared storage resources.

  How can I check the san and the vmfs format is aligned properly?

  Current situation of our under execution SAN:

  the 4 mb block size

  filesystem 3.31

  Also where do I check for bottlenecks and performance problems?

  Double post:

  Re: Sloow SAN performance. How can I confirm the vmfs format?

  André

• TUNING METHODS

  Can someone please tell me in simple language on performance tuning methods. Please do not give online as reference doc.

  904298 wrote:
  Can someone please tell me in simple language on performance tuning methods. Please do not give online as reference doc.

  Watch an action (for example, a SQL statement), determine the duration of this action, identify what part of this time is spent waiting for resources and what part is working actively towards the completion of the action.

  Performance optimization is simply minimize the waiting time for the resources and make more effective, active work in will not affect anything else in the environment.

  In other words: find out why it is slow, came up with some ideas how to speed up, try the ideas and duck when others cry out that their role is now slowing down.

  ... Simple methods: use all the tools available to determine the component/length of time and then use the tool between your ears to determine possible solutions. It is based on experience.

  Published by: Hans Forbrich on March 8, 2012 12:41

  Moreover, the experience begins by reading. Reading specific manuals. :-)

• performance data not there

  Hi all

  Got a weird problem on capturing performance data.

  I use vc2.5 with sql 2005 full.

  My performance measures do not go beyond 24 hours.

  If you look at the chart options and try to choose anything, except the last day, there no data available.

  This host vc has been for more than a year.

  Anyone see this? Looks like some sort of setting but I can't seem to locate one.

  Thank you.

  I saw this, the data that you see are the perf data stored on the local host not in the comic book.  I suggest that you check the stat roller combined employment and ensure that they are running in the DB.  If they don't work I have very long and you have a large number of hosts and VM you're in trouble.  This could take hours or days to run the roll ups.  You can see that the tables affected truncating and just get the work right now to be your only option.  This is a Ko that explains how to solve the jobs:

  KB:1004382

  http://KB.VMware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=1004382&sliceId=1&docTypeID=DT_KB_1_1&dialogID=4558589&StateID=0%200%203759361