| Author |
Message
|
| mqjeff |
 Posted: Thu Aug 18, 2011 4:20 am Post subject: |
 |
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| BBM wrote: |
| By the way, the sending end is on 6.0.2.2 |
You should encourage them to apply fixPacks. |
|
| Back to top |
|
 |
| bruce2359 |
| Posted: Thu Aug 18, 2011 5:02 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9475
Location: US: west coast, almost. Otherwise, enroute.
|
This is a test channel, right? So, few/no messages much of the time?
Compare channel defs that work in production to those of this channel. Any differences (other than channel name, etc.)?
When the channel goes indoubt, what are the msg sequence numbers at both ends? Is the receiving end 0?
_________________
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 |
|
|
| BBM |
| Posted: Thu Aug 18, 2011 5:07 am Post subject: |
|
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
Hi,
Agreed the versions are out of date. This is a test channel and we have no production equivalent as yet, so nothing working to compare it with.
When the channel goes in-doubt the sequence numbers differ by the number of messages that haven't been delivered. so if 5 messages are in transit and the receiving end has a sequence number of 5 the sender will have a sequence number of 10.
I am going to push for upgrades on both sides.
Thanks |
|
| Back to top |
|
|
| gbaddeley |
| Posted: Thu Aug 18, 2011 3:34 pm Post subject: |
|
|
Jedi Knight
Joined: 25 Mar 2003
Posts: 2538
Location: Melbourne, Australia
|
Upgrading to the latest FixPack is always a good thing, but you should scan the list of APARs for each FixPack to see if there are any possible matches on the symptoms before going to the large effort of applying FixPacks, with the possibility that it won't solve your problem at all.
MQ has very good forward and backward version compatibility on channel connections. We have lots MQ 5.3, 6.0, 7.0 and 7.0.1 on various platforms in production and they all talk to each other with perfect reliability.
As soon as you mentioned TCP/IP SACKs, the integrity of the network and TCP stacks comes into question. You need to work hard with your networking team to identify the problem. It might be an intermittently bad NIC card, connector, cable or switch / router. Try to find a way to reliably reproduce the problem.
_________________
Glenn |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Aug 18, 2011 8:04 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20763
Location: LI,NY
|
I seem to remember that there was a bit of a problem between 6.0.2.x and 7.0.1.x before 7.0.1.3... This is why I push for the upgrade. If the upgrade does not fix the problem, involve your network people (and open a PMR?).
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| BBM |
| Posted: Fri Aug 19, 2011 6:07 am Post subject: |
|
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
Good advice indeed - and thanks for everyone's input.
I did open a PMR on the off-chance (and because the issue was high profile) and IBM advised us that there was no known defect with these symptoms for either version of MQ in use.
We got the other side to run some tests on their network and the private link between us. It seems they had some bandwidth capping in place which they removed last night. Today everything works perfectly.
We are still a little confused how bandwidth capping would affect messages >1k but not <1k but are happy that we seem to have a permanent fix.
Thanks again! |
|
| Back to top |
|
|
| exerk |
| Posted: Fri Aug 19, 2011 6:59 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| BBM wrote: |
| We are still a little confused how bandwidth capping would affect messages >1k but not <1k... |
Probably in the same way that ISPs traffic shape...
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Fri Aug 19, 2011 7:35 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9475
Location: US: west coast, almost. Otherwise, enroute.
|
| BBM wrote: |
| We are still a little confused how bandwidth capping would affect messages >1k but not <1k but are happy that we seem to have a permanent fix. |
Did you google "bandwidth capping?"
_________________
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 |
|
|
| BBM |
| Posted: Sat Aug 20, 2011 11:45 am Post subject: |
|
|
Master
Joined: 10 Nov 2005
Posts: 217
Location: London, UK
|
No I didn't Bruce - as happily I already know what bandwidth capping is! Did you?!
I guess I should have qualified why we (myself and my network team) are confused as to why bandwidth capping could affect us.
The bandwidth cap is in place on our corporate private cloud for traffic bursting over and above 10 mbit. To be able to reliably cause the issue by sending messages over 1k and reliably make the issue go away by sending messages under 1k does not point towards a bandwidth cap of 10mbit - or so my very learned network team tell me.
They are trying to establish if anything else was removed when the bandwidth cap was removed.
Still, it's a good thing that the resolution was nothing to do with MQ |
|
| Back to top |
|
|
| exerk |
| Posted: Sat Aug 20, 2011 11:52 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| BBM wrote: |
| They are trying to establish if anything else was removed when the bandwidth cap was removed... |
I believe that bandwidth capping and traffic shaping are different animals, but am more than happy to be put right.
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Sat Aug 20, 2011 12:28 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9475
Location: US: west coast, almost. Otherwise, enroute.
|
| BBM wrote: |
No I didn't Bruce - as happily I already know what bandwidth capping is! Did you?!
The bandwidth cap is in place on our corporate private cloud for traffic bursting over and above 10 mbit. ... or so my very learned network team tell me.
|
There are a variety of ways to cap bandwidth - not limited solely to burst size. It is possible/likely that there may be multiple capping options in place based; and capping can be self-adjusting.
_________________
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 |
|
|
|
|
