| Author |
Message
|
| mrk.for.dev |
 Posted: Thu Mar 18, 2021 2:53 am Post subject: RDQM Hearbeat timeout |
 |
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
Hello,
Is there any way to increase the heartbeat timeout of RDQM?
Ex. Even if the server is out, detect this after 15 secondes |
|
| Back to top |
|
 |
| exerk |
| Posted: Thu Mar 18, 2021 3:13 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
What is the technical/business reason for wanting to increase it?
_________________
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 |
|
|
| mrk.for.dev |
| Posted: Thu Mar 18, 2021 6:00 am Post subject: |
|
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
|
| Back to top |
|
|
| exerk |
| Posted: Thu Mar 18, 2021 6:18 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mrk.for.dev wrote: |
| mini network cut |
I assume from that you mean that part of the network between one or more of the nodes will be temporarily affected, for no more than 15 seconds?
_________________
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 |
|
|
| mrk.for.dev |
| Posted: Thu Mar 18, 2021 7:47 am Post subject: |
|
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
| sometimes the network on the primary server is lost for a few seconds (~3s), during this time the QMs switch to a secondary server. The goal is to increase this timeout to 10 seconds for example, and keep QMs on the primary server even if it is not reachable for less then 10s. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Mar 18, 2021 7:50 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Why the brief network outages? What do your network gurus say about this?
_________________
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: Thu Mar 18, 2021 8:13 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
What he said; fix the network issue rather than mitigate it.
_________________
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 |
|
|
| mrk.for.dev |
| Posted: Thu Mar 18, 2021 8:35 am Post subject: |
|
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
| Nothing special. Difficult to identify the reason. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Thu Mar 18, 2021 9:31 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
| mrk.for.dev wrote: |
| Nothing special. Difficult to identify the reason. |
Huh? Network failures are nothing special and difficult to identify? You need a new network team.
_________________
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 |
|
|
| mrk.for.dev |
| Posted: Fri Mar 19, 2021 2:20 am Post subject: |
|
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
| So there is no way to increase the RDQM timeout? |
|
| Back to top |
|
|
| exerk |
| Posted: Fri Mar 19, 2021 2:30 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mrk.for.dev wrote: |
| So there is no way to increase the RDQM timeout? |
Possibly, but it's not something I have researched. Irrespective of that, try not to use MQ to mitigate problems in other areas, have those areas fix their issues.
_________________
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 Mar 19, 2021 5:17 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
I googled 'rdqm timeout' and the first hit https://www.ibm.com/support/knowledgecenter/en/SSFKSJ_9.0.0/com.ibm.mq.tro.doc/q133450_.htm
I searched ths document for 'timeout' and the first hit was 'Corosync timeout'.
I did a google search for 'Corosync timeout' and found the token needed to set that value.
_________________
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 |
|
|
| mrk.for.dev |
| Posted: Fri Mar 19, 2021 5:45 am Post subject: |
|
|
Novice
Joined: 11 Jan 2021
Posts: 23
|
| That's exactly what I did. I changed the token in totem in /etc/corosync/corosync.conf but I have the impression that it is not taken into account. I set it to 10000=10 seconds. When the primary server is disconnected, the secondary server becomes primary in only 2 seconds. |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Fri Mar 19, 2021 7:19 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9469
Location: US: west coast, almost. Otherwise, enroute.
|
Spend more time reading up on corosync configuration generally, and timeout values specifically.
RDQM is not my specialty. But as I read it, the timeout token is like MQs heartbeat interval, and not a delay feature. Someone should be along shortly.
_________________
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 |
|
|
|
|
