| Author |
Message
|
| sidharth_bora |
 Posted: Sun Jan 04, 2009 11:57 pm Post subject: Disconnect interval time=0 + triggered channel |
 |
|
Voyager
Joined: 24 Nov 2005
Posts: 87
|
Hi All,
we have a MQ set up (not clustered) where we have defined the channel to have DISCINT=0 implying that the channel will be always running whether or not - the messages are arriving on it or not
the same channel has a triggered channel- attribute also enabled
Questions- Can this channel go to inactive state
we received a manual complain to investigate, if the channel was down. Is it that triggering will take priority over DISCNT
if above has some flaws
How can we keep the channel always running (we have resources to keep it running always)
Please can some body throw some light on it.
Cheers
Sid |
|
| Back to top |
|
 |
| sidharth_bora |
| Posted: Sun Jan 04, 2009 11:58 pm Post subject: |
|
|
Voyager
Joined: 24 Nov 2005
Posts: 87
|
just to add on the versions
MQ 5.3 CSD8.0 |
|
| Back to top |
|
|
| Mr Butcher |
| Posted: Mon Jan 05, 2009 12:20 am Post subject: |
|
|
Padawan
Joined: 23 May 2005
Posts: 1716
|
afaik a discint 0 channel will not go inactive. it may go into stopped state on certain error conditions, when MQ decides retry is not useful. in that case triggering will not help (a stopped channel will not be started by triggering).
if you want to make sure your channel is always running, then you should
- make sure it is started at qmgr start
- use monitoring / alerting software to monitor the desired channel status
channel triggering should not be needed in that case, but it will work on inactive channels, so there is no reason why it should not be used additionally.....
_________________
Regards, Butcher |
|
| Back to top |
|
|
| sidharth_bora |
| Posted: Mon Jan 05, 2009 2:02 am Post subject: |
|
|
Voyager
Joined: 24 Nov 2005
Posts: 87
|
thanks for your reply.
If DISCNT interval =0 then it shouldnt go to inactive state, though I agree it can go to stopped state.
I confirm, the channel went to inactive state for the above stated setting
can it be a bug with 5.3
Cheers
Sid |
|
| Back to top |
|
|
| Vitor |
| Posted: Mon Jan 05, 2009 2:48 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sidharth_bora wrote: |
can it be a bug with 5.3
|
Could be - that's a very old level of v5.3; move to CSD 11 or better still move to v6.
IHMO the effects of combining DISCON=0 with triggering are a bit undefined; I'd be interested in the opinions of others here. I would either have a long running channel or a triggered channel, not both.
My personal preference is triggered channels, as I find them less hassle.
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Mon Jan 05, 2009 5:01 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7716
|
You should almost never use DISCINT = 0. Its been discussed here before. Search. "I always want my channels running" is NOT a valid reason.
Even with DISCINT = 0, you will still want the channels set up for triggering, because after the QM restarts, all the channels will come up as Inactive.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| kevinf2349 |
| Posted: Mon Jan 05, 2009 12:49 pm Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
Maybe somebody manually stopped the channel and also manually set the status to inactive in order to allow triggering to restart it?
Have you checked the logs to see what caused it?
...and I fully agree with Peter about DISCINT=0  |
|
| Back to top |
|
|
| sidharth_bora |
| Posted: Mon Jan 05, 2009 6:38 pm Post subject: |
|
|
Voyager
Joined: 24 Nov 2005
Posts: 87
|
Hi, All
thanks for your suggestion and advise. I have checked the logs and there is nothing as such "channel ended normally" or abnormally.
DISCNT=0 is the policy on the prod host - where we have abundant resource & so we have gone ahead with this set up.
we are planning to upgrade to 6.0 - hopefully we shouldnt have this issue.
Peter- you are right, after QM restart- we need the channels to start automatically that is one of the reasons - we have triggered channels in place.
Also if some body could help to know - if receiver channel has some attribute needs to be fixed here as i dont see DISCNT attribute for receiver channel
PS: Before posting in, I searched the forum, but didnt have the requisite explanation / answer to my issue
cheers
Sid |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 05, 2009 7:42 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9406
Location: US: west coast, almost. Otherwise, enroute.
|
Receiver channels are started by the listener.
Please search for more posts on this well-discussed subject.
A false assumption: Just because a channel in in RUNNING state doesn't mean it can send/receive messages.
Briefly, RUNNING status means that the channel is currently sending/receiving messages; OR, if it is not sending/receiving messages, the MCA believes that it can. It is possible that the next time the MCA actually tries to send a message, the channel might fail.
So, it seems that a benefit of a forever-running channel might be realized if messages arrive in the xmit queue sporadically enough that triggering delays the channel start and message delivery. Of course, the channel could fail to start after being triggered for the same reasons that an apparently RUNNING channel can not actually send a message. Darn networks.
| Quote: |
| ...where we have abundant resource |
Historically, disconnect interval allowed MCAs to close queues, freeing up buffers, channel status table entries and processor cycles. With more robust servers, this need has diminished.
_________________
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 |
|
|
| kevinf2349 |
| Posted: Tue Jan 06, 2009 5:55 am Post subject: |
|
|
Grand Master
Joined: 28 Feb 2003
Posts: 1311
Location: USA
|
| sidharth_bora wrote: |
DISCNT=0 is the policy on the prod host - where we have abundant resource & so we have gone ahead with this set up.
|
It isn't just about abundant resources. It is also about security, resilience and recoverability. |
|
| Back to top |
|
|
| Vitor |
| Posted: Tue Jan 06, 2009 6:22 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
| sidharth_bora wrote: |
| DISCNT=0 is the policy on the prod host - where we have abundant resource & so we have gone ahead with this set up. |
It's an odd policy and (in your place) I'd question it. What does abundant resource have to do with it?
| sidharth_bora wrote: |
| we are planning to upgrade to 6.0 - hopefully we shouldnt have this issue. |
Hope is a bad support strategy for a production system. Though I've noticed many sites favour it. 
_________________
Honesty is the best policy.
Insanity is the best defence. |
|
| Back to top |
|
|
|
|
