Forum Discussion
CANSTAYN569
Nimbostratus
Nimbostratus3900 Goes offline
Hi all,
I have a 3900 goes offline after a short while it is online.
Here is the story, i was creating a cluster between 3900 and 6900. I actually did. When i sync 6900 to 3900 ,3900 went offline.
Since 3900 had a config for serial failover , i thought this might be the reason and disabled serial failover.
I've followed the procedure below :
Log in to the tmsh utility by typing the following command: tmsh
Modify the failover.usetty01 database key to enable by typing the following command: modify sys db failover.usetty01 value enable
Note: By default, the failover.usetty01 database setting synchronizes between units in a high-availability pair during a ConfigSync operation.
Save the change by typing the following command: save /sys config
To restart the sod process, type the following command: restart sys service sod
After that the device became after for few seconds. And went offline again.
It becomes active then it says :
Oct 18 18:42:39 DISF5_1 emerg logger: Re-starting bigd
I have this in ltm :
Oct 18 18:42:39 DISF5_1 info bigd.start[13298]: Execing bigd: /usr/bin/bigd bigd -S
Any ideas why am i facing this offline issue?
Regards Sadik
CANSTAYN569I have this log under ltm. But i can not understand why the unit fails over.
Oct 18 18:49:11 DISF5_1 warning sod[14832]: 01140029:4: HA proc_running bd fails action is go offline and down links.
Oct 18 18:49:11 DISF5_1 notice sod[14832]: 010c0050:5: Sod requests links down.
Oct 18 18:49:11 DISF5_1 notice sod[14832]: 010c0054:5: Offline for traffic group /Common/traffic-group-1.
Oct 18 18:49:11 DISF5_1 notice sod[14832]: 010c003e:5: Offline
Oct 18 18:49:11 DISF5_1 notice tmm3[10684]: 01340011:5: HA unit 1 state change: from 1 to 0.
Oct 18 18:49:11 DISF5_1 notice tmm1[10684]: 01340011:5: HA unit 1 state change: from 1 to 0.
Oct 18 18:49:11 DISF5_1 notice tmm2[10684]: 01340011:5: HA unit 1 state change: from 1 to 0.
Oct 18 18:49:11 DISF5_1 notice tmm[10684]: 01340011:5: HA unit 1 state change: f rom 1 to 0.
Oct 18 18:49:11 DISF5_1 err mcpd[4746]: 01070069:3: Subscription not found in mc pd for subscriber Id BIGD_Subscriber.
Oct 18 18:49:11 DISF5_1 info lacpd[4744]: 01160016:6: Failover event detected. ( Switchboard failsafe disabled while offline)
Oct 18 18:49:11 DISF5_1 err bcm56xxd[6215]: 012c0010:3: Failover event detected. Marking external interfaces down. bsx.c(3988)
And here is the evidence that it goes active just for a while :
[root@DISF5_1:Offline:Standalone] log tmsh restart sys service sod [root@DISF5_1:Offline:Standalone] log [root@DISF5_1:Active:Standalone] log [root@DISF5_1:Active:Standalone] log
Vijay_ECirrus
Clustering across different platforms isn't supported by F5. If you can make it work, great. If you can't, F5 isn't going to help out if you raise a case.
- CANSTAYN569
Another key point, the 6900's license is expired. But it is functional. It does not welcome me with the re-activate page. But still its not licensed. So i have synced from an unlicensed device, then the 3900 went offline.
- Vijay_E
The expired license could be the problem.
- CANSTAYN569
Interesting updates. I have licensed the 6900. But its license did not include ASM. so i disabled asm on both unit. It worked perfectly. Then i have switched the license back to old one (which includes ASM) no problem at all again. When i enabled asm, BOOM! 3900 went offline. Same problem. I have just disabled the ASM and it came back online, cluster is working fine again.
Note : ASM was enabled at the very beginning and when the device was stand alone.
I will try to get a valid license to 6900 which includes ASM and then re-provision that module and see what happens.
Regards