Support and News Forum

Hi all,

I have a Problem with the new V77:

Randomly my vSRXes start to freeze - completely.
Even when clicking "stop" it still thinks it's running (but i verified with top, that it stopped).
In this case i have to reboot the whole eve host to have everything "offline" and to start my lab again - this is really annoying - has anyone else experienced this problem?
I'm using vSRX 15.1D100 and vSRX 17.3R1 with EVE V77 and virtioa.qcow2 disks.
I labbed a lot with v71 and never had problems like this.

Regards
Chris

Could be related to the new per node cpu limiting which was added in v77. You can try disable by setting the value to 1 in the node config.

CPU limit can be disabled per node from web UI, edit node and uncheck CPU limit

I disabled it globally and set it only on the "per node" basis - seems to work now.
But as soon as i enable it globally AND on the node (like double limiting) the fun begins...

CPU limit on globally will monitor all Qemu process Except those not checked.

For all Juniper, it seems mandatory to uncheck cpulimit.
Indeed, Juniper doesn't support pause mechanism. The engine for packet processing require to poll the cpu without interruption...
I don't like much this model of event handling but that's the Juniper choice....

E.

i'm seeing this same issue with IOL instances as well

I'm also seeing the same issue with IOL nodes and I'm able to recreate it. When the terminal session times out and you restart it, then type a command like show eigrp tech support the node freezes but then unfreezes after about 20 minutes.



After it unfreezes it comes up with these errors but the nodes work fine.

*Sep 12 22:42:08.505: %AMDP2_FE-6-EXCESSCOLL: Ethernet0/2 TDR=0, TRC=0
*Sep 12 22:42:38.517: %AMDP2_FE-6-EXCESSCOLL: Ethernet0/2 TDR=0, TRC=0
*Sep 12 22:43:08.525: %AMDP2_FE-6-EXCESSCOLL: Ethernet0/2 TDR=0, TRC=0
*Sep 12 22:43:38.529: %AMDP2_FE-6-EXCESSCOLL: Ethernet0/2 TDR=0, TRC=0


Is there an option or command to manual force nodes to restart. (shutdown, start do not work when the node is frozen)

I had the same problem - either wait (30mins in my case) or reboot the eve-host - the Webinterface does not recognize, that the qemu isn't running anymore - it still shows my device as "up" even when all qemu instances are down - really weird...

Today i found a workaround for me:

More Actions > Console to all nodes (native mode) unfreezes everything for me - strange, right?

my connection is native always and it has been doing this since the upgrade to 77

Same here, I've seeing IOL and VPC node freezing.