| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| RDQM and DRBD installation on 9.2 and Rhel 8 |
« View previous topic :: View next topic » |
| Author |
Message
|
| RDQMAN |
 Posted: Mon Oct 16, 2023 8:34 am Post subject: RDQM and DRBD installation on 9.2 and Rhel 8 |
 |
|
Newbie
Joined: 16 Oct 2023
Posts: 4
|
Hi all,
In HA/DR configuration I have a problem with first connection between first and secondary DR istances, DR status remains in "Waiting for initial connection":
this is an rdqmstatus of my situation on first node:
[mqm@sd80aciwmq01 ~]$ rdqmstatus -m ITACMQ01
Node: sd80aciwmq01.pippo.it
Queue manager status: Running
CPU: 0.05%
Memory: 135MB
Queue manager file system: 97MB used, 9.8GB allocated [1%]
HA role: Primary
HA status: Normal
HA control: Enabled
HA current location: This node
HA preferred location: This node
HA floating IP interface: None
HA floating IP address: None
DR role: Primary
DR status: Waiting for initial connection
DR port: 7002
DR local IP address: 10.113.170.20
DR remote IP address list: 10.113.170.23,10.113.170.24,10.113.170.25
DR current remote IP address: Unknown
Node: sd80aciwmq02.pippo.it
HA status: Normal
Node: sd80aciwmq03.pippo.it
HA status: Normal
and in the second node:
[mqm@sd80aciwmq04 ~]$ rdqmstatus -m ITACMQ01
Node: sd80aciwmq04.pippo.it
Queue manager status: Ended immediately
HA role: Primary
HA status: Normal
HA control: Enabled
HA current location: This node
HA preferred location: This node
HA floating IP interface: None
HA floating IP address: None
DR role: Secondary
DR status: Waiting for initial connection
DR port: 7002
DR local IP address: 10.113.170.23
DR remote IP address list: 10.113.170.20,10.113.170.21,10.113.170.22
DR current remote IP address: Unknown
Node: sd80aciwmq05.pippo.it
HA status: Normal
Node: sd80aciwmq06.pippo.it
HA status: Normal
DR disk resource is stucked in inconsistent state:
drbdadm status on primary DR node
itacmq01 role:Primary
disk:UpToDate
sd80aciwmq02.pippo.it role:Secondary
peer-disk:UpToDate
sd80aciwmq03.pippo.it role:Secondary
peer-disk:UpToDate
itacmq01.dr role:Primary
disk:UpToDate
_remote connection:StandAlone
drbdadm status on secondary DR node
itacmq01 role:Primary
disk:UpToDate
sd80aciwmq05.pippo.it role:Secondary
peer-disk:UpToDate
sd80aciwmq06.pippo.it role:Secondary
peer-disk:UpToDate
itacmq01.dr role:Secondary
disk:Inconsistent
_remote connection:Connecting
It seems to not have problems on network firewall communication, ssh key exchange, no selinux config (disabled), or volume group problems.
What could it be?
Do you have any suggestions?
Thank you very much. |
|
| Back to top |
|
 |
| hughson |
| Posted: Tue Oct 17, 2023 2:33 am Post subject: |
|
|
Padawan
Joined: 09 May 2013
Posts: 1967
Location: Bay of Plenty, New Zealand
|
|
| Back to top |
|
|
| RDQMAN |
| Posted: Tue Oct 17, 2023 4:56 am Post subject: |
|
|
Newbie
Joined: 16 Oct 2023
Posts: 4
|
Thank you for your reply 
Sure, I've used this rdqm.ini configuration:
In the Primary 3 HA/DR nodes:
#Primary HA/DR 3 servers:
Node:
Name=sd80aciwmq01.pippo.it
HA_Primary=10.113.170.36
HA_Replication=10.113.170.20
# DR_Replication=10.113.170.20
Node:
Name=sd80aciwmq02.pippo.it
HA_Primary=10.113.170.37
HA_Replication=10.113.170.21
# DR_Replication=10.113.170.21
Node:
Name=sd80aciwmq03.pippo.it
HA_Primary=10.113.170.38
HA_Replication=10.113.170.22
# DR_Replication=10.113.170.22
#DRGroup:
# Name=DR1
# DR_Replication=10.113.170.23
# DR_Replication=10.113.170.24
# DR_Replication=10.113.170.25
In three Secondary HA/DR nodes (4-5-6)
#Secondary HA/DR 3 servers:
Node:
Name=sd80aciwmq04.pippo.it
HA_Primary=10.113.170.39
HA_Replication=10.113.170.23
# DR_Replication=10.113.170.23
Node:
Name=sd80aciwmq05.pippo.it
HA_Primary=10.113.170.40
HA_Replication=10.113.170.24
# DR_Replication=10.113.170.24
Node:
Name=sd80aciwmq06.pippo.it
HA_Primary=10.113.170.41
HA_Replication=10.113.170.25
# DR_Replication=10.113.170.25
#DRGroup:
# Name=DR1
# DR_Replication=10.113.170.20
# DR_Replication=10.113.170.21
# DR_Replication=10.113.170.22
DR is commented, I use these string to crtmqm:
Step 1: Node 2-3
crtmqm -sxs -rr p -rl 10.113.170.20,10.113.170.21,10.113.170.22 -ri 10.113.170.23,10.113.170.24,10.113.170.25 -rp 7001 -fs 10200M ITACMQ01
Step 2: Node 1
crtmqm -sx -rr p -rl 10.113.170.20,10.113.170.21,10.113.170.22 -ri 10.113.170.23,10.113.170.24,10.113.170.25 -rp 7001 -lp 3 -ls 2 -lf 8192 -lln -fs 10200M ITACMQ01
Step 3: Node 5-6
crtmqm -sxs -rr s -rl 10.113.170.23,10.113.170.24,10.113.170.25 -ri 10.113.170.20,10.113.170.21,10.113.170.22 -rp 7001 -fs 10200M ITACMQ01
Step 4: Node 4
crtmqm -sx -rr s -rl 10.113.170.23,10.113.170.24,10.113.170.25 -ri 10.113.170.20,10.113.170.21,10.113.170.22 -rp 7001 -fs 10200M ITACMQ01 |
|
| Back to top |
|
|
| RobertK |
| Posted: Tue Oct 17, 2023 9:05 am Post subject: IP subnet configuration |
|
|
Newbie
Joined: 17 Oct 2023
Posts: 4
Location: NC, USA
|
Hi RDQMAN,
It is not a good practice to use IP addresses on the same subnet assigned to the same node (even if on different interfaces). RDQM uses totem membership tokens (corosync) and DRBD pings (DRBD) to verify cluster membership status and replication availability.
If a ping goes out on interface ens192, a response is not guaranteed to be returned to the same interface. This will cause cluster instability and replication problems or slowness.
I recommend placing DRBD replication interfaces for all the nodes on a single subnet, corosync totem on a different subnet, and DR replication interfaces on another subnet.
Only as an example, see the rdqm.ini file below:
| Code: |
Node:
Name=sd80aciwmq01.pippo.it
HA_Primary=10.113.170.36
HA_Replication=10.113.171.20
# DR_Replication=10.113.172.20
Node:
Name=sd80aciwmq02.pippo.it
HA_Primary=10.113.170.37
HA_Replication=10.113.171.21
# DR_Replication=10.113.172.21
Node:
Name=sd80aciwmq03.pippo.it
HA_Primary=10.113.170.38
HA_Replication=10.113.171.22
# DR_Replication=10.113.172.22
#DRGroup:
# Name=DR1
# DR_Replication=10.113.173.23
# DR_Replication=10.113.173.24
# DR_Replication=10.113.173.25 |
If you don't define a DR replication interface in the rdqm.ini file, DR Replication will happen on the HA_Replication interface.
I hope this helps! |
|
| Back to top |
|
|
| RDQMAN |
| Posted: Wed Oct 18, 2023 8:02 am Post subject: |
|
|
Newbie
Joined: 16 Oct 2023
Posts: 4
|
Hi RobertK,
thank you for your reply and for your opinion.
So we have 3 vlans on same subnet 10.113.170.*
eth0 subnet for data - 10.113.170.4-5-6-7-8-9
eth1 for drbd - 10.113.170.20-21-22-23-24-25
eth2 for pacemaker - 10.113.170.36-37-38-39-40-41
and 2 Floating IP 10.113.170.12 and 10.113.170.13
So there'is no way to get it going stable with this vlan conf?
Thank you very much! |
|
| Back to top |
|
|
| RobertK |
| Posted: Wed Oct 18, 2023 8:38 am Post subject: Network cross-talk |
|
|
Newbie
Joined: 17 Oct 2023
Posts: 4
Location: NC, USA
|
RDQMAN:
Question1:
DRBD interface pings other DRBD interfaces on other RDQM nodes to ensure they are all still alive and running as part of its heartbeat-like checks. Are there measures in place that guarantee the ping request from the replication port on nodex gets delivered to the replication interface on nodey and nodez and not the pacemaker or floating IP address interfaces?
If no, it is not good practice to use same-subnet-IPs because drbd ping responses may not be from the expected drbd replication interface.
I am not a networking expert and do not understand your network configs, so I am unable to comment on whether or not the communication will properly work given all of the IPs are on the same subnet.
You could reduce the number of interfaces so pacemaker, drbd replication, and floating IPs all use the same subnet. To do this, however; you will need to delete the queue manager(s), delete the HA group, only define the HA_Replication interface in the rdqm.ini file and recreate the HA cluster and queue manager(s).
Does this help? |
|
| Back to top |
|
|
| RDQMAN |
| Posted: Thu Oct 19, 2023 6:35 am Post subject: |
|
|
Newbie
Joined: 16 Oct 2023
Posts: 4
|
Hi RobertK,
thank for reply, yes it helps
I will trying and I'll update you. |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
