| Author |
Message
|
| sharonv |
 Posted: Tue Feb 11, 2014 12:56 am Post subject: |
 |
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
HI
I'm using sender reviever channels for connecting ( system A ) to my MQ i also have a set of sender receiver that connect from my MQ to the system A
I dont have any issue with my sender ,
only my reciever get out of sync
and it defently happens after the system A sender , reached the disconeect interval and go down , when going up the reciever sequence is higer then the sender.
BTW
i have several systems connecting to my MQ and this havior is happening in all of them
i have another linux MQ server which works ok with this systems
any idea's were to look ?? |
|
| Back to top |
|
 |
| PeterPotkay |
| Posted: Tue Feb 11, 2014 5:18 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
| PeterPotkay wrote: |
| What are the details? Always off by one? Or, the sending side goes back to 1? Or the sending side stays correct but the receiving side goes up some random amount? |
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 5:42 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
HI
Two senarios , the first when disconnect interval happens on the system A side , then there is a jump of one or two
if i stop my qmgr and start it them i see a random number sequence number |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Feb 11, 2014 5:52 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sharonv wrote: |
| ...i see a random number sequence number |
There's no such thing! WMQ does not randomly assign numbers to a sequence counter. I think you need to trap all incoming connections to your channels and examine whether they're coming from a single entity, or multiple entities, or whether someone has stuffed an F5 or similar between you and the sending end.
_________________
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 |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 6:14 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
HI
i agreee , its not random but with a bigger difference then one or two,
one of the connecting system is under my responsibilities
it connect to basic sender /receiver channels
i have a remote queue from this system to my MQ
i do a simple test
i start the sender on the remote system ( no application is connected to
it ) for test purpose i set its disconnect time to 1 minute
i see the reciever starts on my MQ system
i put a message through the remote queue , and it arrives to my MQ systme ( all queues are persistend btw )
i then wait for one minute , the sender goes down to inactive as expected and so his my reciever
i send a message again on the remote system the sender awakes and goes into retry
i check the log and i see the message sequence number problem |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Feb 11, 2014 6:25 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sharonv wrote: |
| ...i check the log and i see the message sequence number problem |
And just to be sure, you're finding that in the log of the sending system? What are the exact numbers you're seeing at each end please?
_________________
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 |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 6:54 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
from the sender log
a message with sequence number 19 has been sent when sequence number 16 was expected
i did another test
i created two new QMs one on the sender machine and one on my MQ machine
i connected them with sender reciver , repeates the scenario and it works allright
the only difference between the new qm and the other ones in my system
is that the new one is not connected to the veritas cluster resource
Do you have any other idea's ? |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Feb 11, 2014 7:03 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9400
Location: US: west coast, almost. Otherwise, enroute.
|
| sharonv wrote: |
from the sender log
a message with sequence number 19 has been sent when sequence number 16 was expected |
In your test, did you send 3 messages? (the difference between 16 and 19.)
_________________
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 |
|
|
| exerk |
| Posted: Tue Feb 11, 2014 7:09 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
I've split your post off from the original that's three years old.
| sharonv wrote: |
from the sender log
a message with sequence number 19 has been sent when sequence number 16 was expected |
Which implies the sending end has sent an additional three messages to an instance of a RCVR.
| sharonv wrote: |
i did another test
i created two new QMs one on the sender machine and one on my MQ machine
i connected them with sender reciver , repeates the scenario and it works allright
the only difference between the new qm and the other ones in my system
is that the new one is not connected to the veritas cluster resource
Do you have any other idea's ? |
This is the first mention of a High-Availability set up, which would have been useful to know up-front. Please describe in detail the set-up, and whether this is a recent occurrence, i.e. everything worked in the past, or whether this problem has been apparent from the beginning.
_________________
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 |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 7:36 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
First thanks for your willing to assist
More details
my mq system was previously running on windows 2003 server
a week ago i moved it to a linux server and then we started getting this issue
on the windows it was with MQSC
on the linux with Verital ( active passive , shared disk for configuration and logs)
regarding my issue
just to make sure i disconnecte one of the qmgrs with a problem from the cluster
and the issue still exists
i use also blockip and tried to remove it as well but the issue still exists
( i have another linux mq server with the same definition that works ok with the veritas and the block ip
did all comparing i can and it looks the same
i'm thinking about recreating the qmgrs .... any ideas? |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Feb 11, 2014 7:49 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sharonv wrote: |
regarding my issue
just to make sure i disconnecte one of the qmgrs with a problem from the cluster |
Just to be sure, you removed 'one of the qmgrs' from Veritas Cluster control? We're not talking about a queue manager cluster?
_________________
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 |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 7:55 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
No , its not mq cluster
its two instances of MQ installed on two seperated machine with shared disk for log and configuration running active passive
i just now deleted my reciever channel and recreatred it and the problem was resloved ....
I'm planning to do the same on all other qmgrs
one more thing as it comes to mind that about a month ago i did an upgrade from 7.0.1.8 to 7.0.1.11 , this was before this computer was in testing , could be the upgrade didnt go well , and creating the object solved the issue ??
just before i do the delete any ideas? |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Feb 11, 2014 8:18 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| sharonv wrote: |
No , its not mq cluster
its two instances of MQ installed on two seperated machine with shared disk for log and configuration running active passive |
So what exactly do you mean by 'removed one of the qmgrs'?
| sharonv wrote: |
i just now deleted my reciever channel and recreatred it and the problem was resloved ....
I'm planning to do the same on all other qmgrs |
So, problem solved but root cause still to be established.
| sharonv wrote: |
| one more thing as it comes to mind that about a month ago i did an upgrade from 7.0.1.8 to 7.0.1.11 , this was before this computer was in testing , could be the upgrade didnt go well , and creating the object solved the issue ?? |
So did the issue only arise after the upgrade? Was there no functional testing carried out to check that what function you had when it was running on Windows was also replicated in the move to Linux? Did you not initiate a fail-over while message traffic was going across the line to check channel recovery?
_________________
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 |
|
|
| JosephGramig |
| Posted: Tue Feb 11, 2014 8:21 am Post subject: |
|
|
Grand Master
Joined: 09 Feb 2006
Posts: 1231
Location: Gold Coast of Florida, USA
|
| Please post your qm.ini file. |
|
| Back to top |
|
|
| sharonv |
| Posted: Tue Feb 11, 2014 8:43 am Post subject: |
|
|
Apprentice
Joined: 03 Feb 2011
Posts: 27
|
HI
"Removed one of the qmgr " meaning deleting its cluster resource , so it is not managed by the cluster any more
root cuase not established , yet i'm planning to recreate the entire qmgrs from scratch
my regrets is that i prepared this machine about 6 monthes ago and only a week ago i could actually test it with live client ( i;m still in a
test environment )
till then i just checked mq was running and that i could connect with it using mqmon and application . but not with remote systems , it even was tested with the cluster and failover but not with real remote clients using sender reciever channels
and ... from security issues i cant put the qm.ini file online , well i could but i need to delete all qm names and ...... |
|
| Back to top |
|
|
|
|
