VMware Cloud Community
racom
Enthusiast
Enthusiast

VDP: heavy traffic on lo

Hi,

I successfully restored one VM last Thursday. I'm checking CPU Usage of VDP now and it increased to three times since then. Two java processes run constantly and I can see heavy traffic on lo (25 Mb/s along to Zenoss monitoring). Is it OK? Could I wait? Or reboot?

Thanks for advice

Tomas

Reply
0 Kudos
3 Replies
racom
Enthusiast
Enthusiast

I've got virtual machine CPU usage alarm for VDP this morning. It increased from cca 55% to 95% (there were 15% before VM restoring). Traffic on lo increased to almost 40Mb/s.

I've checked /var/log/messages on VDP and there were a lot of "scsi: killing requests for dead queue" messages around VM restoring and plenty of ssh connections for root from 127.0.0.1 (once in every minute) as well as for admin from eth0 IP (usually four times in a minute) since then - contrary to 2 + 5 connections once a day at 8 AM before. What is going on?

Everi night one to three backup jobs for different VMs fail since then in the adition (no more info in VDP reports). I've set 7 backup jobs for a toatl of 24 VMs.

Reply
0 Kudos
racom
Enthusiast
Enthusiast

It's a known issue - http://kb.vmware.com/kb/2041068. Soft restart of VDP helped.

Reply
0 Kudos
racom
Enthusiast
Enthusiast

The same problem again. CPU usage was solved after reboot but heavy traffic on lo persists. I can see RX/TX about 18 Gb after 1 hour uptime. Most of the traffic is to gsan or java/avamar process. What's the reason for using localhost device this way?

Reply
0 Kudos