| Author |
Message
|
| zpat |
 Posted: Sat Mar 08, 2014 9:43 am Post subject: |
 |
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
Run the "mqconfig" script to check you have the correct kernel settings.
http://www-01.ibm.com/support/docview.wss?uid=swg21271236
_________________
Well, I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error. |
|
| Back to top |
|
 |
| romankhar |
| Posted: Sat Mar 08, 2014 12:02 pm Post subject: |
|
|
Novice
Joined: 23 Jan 2014
Posts: 12
|
| bruce2359 wrote: |
| romankhar wrote: |
| I will work on persistent messaging tests as it does not need as much bandwidth. |
Really? A 100Meg persistent message uses less bandwidth than a 100Meg non-persistent message? |
Very funny 
The messages are the same size, but despite the fact that I have 4 SSDs in the server, the rate of 1MB persistent messages is about half of the rate of non-persistent messages, so the network does not get saturated.
I will package results and publish on my blog when all is ready - I will also include Apache ActiveMQ results: http://whywebsphere.com |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Mar 08, 2014 1:03 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| romankhar wrote: |
| ...the rate of 1MB persistent messages is about half of the rate of non-persistent messages ... |
Rate of what exactly? What exactly are you measuring?
Are you measuring only network transmission? Network bandwidth is measured in Mb or Gb per second. Are you saying that persistent messages take more Mb or Gb per second to flow across the network than non-persistent messages?
Are you including the time it takes for your application to create the messages? Are you saying that the aggregate throughput is less for persistent messages?
Please be precise.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| romankhar |
| Posted: Sat Mar 08, 2014 2:01 pm Post subject: |
|
|
Novice
Joined: 23 Jan 2014
Posts: 12
|
| bruce2359 wrote: |
| romankhar wrote: |
| ...the rate of 1MB persistent messages is about half of the rate of non-persistent messages ... |
Rate of what exactly? What exactly are you measuring?
Are you measuring only network transmission? Network bandwidth is measured in Mb or Gb per second. Are you saying that persistent messages take more Mb or Gb per second to flow across the network than non-persistent messages?
Are you including the time it takes for your application to create the messages? Are you saying that the aggregate throughput is less for persistent messages?
Please be precise. |
I am measuring number of messages per second for the REQUESTOR which sends the message to the server and later gets a reply back from RESPONDER. Here is a description of my tests for those interested - I will update it a bit and publish on the blog, but this is pretty close to final: https://docs.google.com/document/d/1IL_xFkXVjgkIqfUI7HVloOE9iees3qL1YHIls9BxWUI/edit?usp=sharing
Since persistent messaging is slower, the load on the network is less.
Here are preliminary performance results - still in the process of tuning things for both WMQ and for AMQ: https://docs.google.com/spreadsheets/d/1HT-HuSybds9UE40a9pSFVn3sTAJBJvKrA5BGolKQHTA/edit?usp=sharing |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Mar 08, 2014 2:51 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
|
| Back to top |
|
|
| fjb_saper |
| Posted: Sat Mar 08, 2014 8:56 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Moving to performance forum
By the way I was able to access both links. Nice document on the results! 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| romankhar |
| Posted: Tue Mar 18, 2014 5:28 am Post subject: |
|
|
Novice
Joined: 23 Jan 2014
Posts: 12
|
|
| Back to top |
|
|
| romankhar |
| Posted: Tue Mar 18, 2014 5:32 am Post subject: |
|
|
Novice
Joined: 23 Jan 2014
Posts: 12
|
I always run mqconfig before and after WMQ install - here is what I have now:
| Code: |
[roman@mqhost mq]$ /opt/mqm/bin/mqconfig
mqconfig: Analyzing Red Hat Enterprise Linux Server release 6.5 (Santiago)
settings for WebSphere MQ V7.5
System V Semaphores
semmsl (sem:1) 1000 semaphores IBM>=500 PASS
semmns (sem:2) 30 of 512000 semaphores (0%) IBM>=256000 PASS
semopm (sem:3) 500 operations IBM>=250 PASS
semmni (sem:4) 9 of 2048 sets (0%) IBM>=1024 PASS
System V Shared Memory
shmmax 68719476736 bytes IBM>=268435456 PASS
shmmni 58 of 4096 sets (1%) IBM>=4096 PASS
shmall 219480 of 4294967296 pages (0%) IBM>=2097152 PASS
System Settings
file-max 6976 of 1000000 files (0%) IBM>=524288 PASS
tcp_keepalive_time 300 seconds IBM<=300 PASS
Current User Limits (roman)
nofile (-Hn) 20480 files IBM>=10240 PASS
nofile (-Sn) 20480 files IBM>=10240 PASS
nproc (-Hu) 58 of 31285 processes (0%) IBM>=4096 PASS
nproc (-Su) 58 of 4096 processes (1%) IBM>=4096 PASS
|
|
|
| Back to top |
|
|
|
|
