Ah, this time it actually reported a bug:
[Wed May 29 11:49:24 2019] perf: interrupt took too long (2544 > 2500),
lowering kernel.perf_event_max_sample_rate to 78500
[Wed May 29 11:58:41 2019] mlx5_core :01:00.0 ib0: "orted" wants to know my
dev_id. Should it look at dev_port instead? See
Somewhat different the second time around. This is with the 5.0.05-15
kernel.
[Tue May 28 17:36:56 2019] general protection fault: [#1] SMP PTI
[Tue May 28 17:36:56 2019] CPU: 23 PID: 0 Comm: swapper/23 Not tainted
5.0.0-15-generic #16~18.04.1-Ubuntu
[Tue May 28 17:36:56 2019] Hardware
Sadly it died with this crash:
[Tue May 28 14:46:58 2019] ib0: enabling connected mode will cause multicast
packet drops
[Tue May 28 14:46:58 2019] ib0: mtu > 4092 will cause multicast packet drops.
[Tue May 28 14:46:58 2019] IPv6: ADDRCONF(NETDEV_CHANGE): ib0: link becomes
ready
[Tue May 28
I'm having similar problems with rsync killing a 18.04 head node that's
writing to a NFS share. It would die pretty quickly with rsync or
globus. We had a dozens of crashes, we originally thought it was a
hardware issue, but switching to a different server didn't help.
We tried:
```
$ apt
Fully patched 16.04 (as of Dec 3rd) still has this problem.
None of the above mentioned fixes (lsusb or echo on to
/sys/bus/usb/devices/*/power/control) fixes the problem.
I'm on a fairly new Xeon E3 Gigabyte motherboard with an E3-1230.
My new Pixel XL is a USB 3 device and when plugged into a
5 matches
Mail list logo
