we've been testing VMX v14.1 in the lab without any problems. We're now doing some tests on the same hosts with v15.1 but we're seeing some issues with the vfp:
RUNTIME: Logical core 1 (I/O) main loop.
RUNTIME: Logical core 2 (worker 0) main loop.
traps: vmxt trap divide error ip:f7620044 sp:ff9ac5d4 error:0 in libgcc_s.so.1[f7618000+1b000]mpc : 0Total cores: core.vmxt.1447189176.1000.gzCore: core.vmxt.1447189176.1000.gz/usr/share/pfe/start_vmxt.sh: line 7: 1000 Floating point exception(core dumped) /usr/share/pfe/vmxt -Nvar: core.vmxt.mpc.1447189176.1000.gzcore.vmxt.mpc.1447189176.1000.gz file sent
vpf doens't come up so no interface to be seen on the vcp.
kernel: $ 3.19.0-25-genericlibvirt: $ /usr/sbin/libvirtd (libvirt) 1.2.8
root@vfp-vmx1:~# freetotal used free shared buffers cachedMem: 7791248 4538568 3252680 0 12584 121756-/+ buffers/cache: 4404228 3387020Swap: 0 0 0
The weird thing is, after rebooting the vfp it sometimes works?????
When it keeps on running, interfaces are showing up in the vcp and everything working fine.
But most of the time, we're getting that core dump and riot doesn't start.
Has anyone seen this before?
I'd suggest you open a JTAC case and provide core dump files for further investigation.
took me some time but core dump is located on the vcp, not on the vfp:
root@% pwd/var/tmproot@% ls -l core.vmxt*-rw-r--r-- 1 root field 2452908 Nov 11 17:44 core.vmxt.mpc.1447263857.1065.gz
there is no support contract on this vmx, just trying it out in the lab, let's see if I can create a case.
bummer, can't open a case, no support contract, duhh.
I will get in contact with our local juniper rep. let's see if he can can do anything with it.
For anyone from the vmx team that want's to have a look, core dump is attached. (.gz renamed in .zip)