Hi,
I was seeing in vROps and found that one of the machine is dropping receiving packets as shown below and I checked with user and he did not noticed anything.
can anyone is seeing this? is there any reason why thous packets are dropping?
Thank you,
Pkmr.
Asking a user to confirm if they are "noticing anything" is not a way to validate the existence of dropped packets. You can check the counter inside vCenter for this VM to corroborate. Also know that in the past, there have been some effected builds of ESXi which erroneously report dropped packets. I don't remember the versions, but a search would turn up more.
sure I will cross check in vCenter too .
dropping packets should impact the user experience right?
Thank you.
I think it may or may not because here as previous command saying few ESX builds had this issue
however check with you build with affected ESX builds, what build are you using? so that some one can tell you either it is in affected or not
Regards
I am using ESX 6.0 U1, is it under affected list?
Thank you
This issue is resolved in vROPs 6.0.1. Please refer to this KB 2105109
are you using NSX dfw in your environment? i noticed that in more recent versions of vRops and vCener / NSX when we switched on the dfw i started to see the dropped packets increase (nothing to worry about in my case as it was expected)
Yeah I am using NSX DFW,
yes, user is not experiencing any issues, so it is expected and is there any workaround for it?
Regards,
Pkmr.
right... so I would say its the dfw blocking stuff that is contributing to the dropped packets metric.
if you want to confirm this add a server with dropping packets to the exception list in NSX manager and see if the dropped packets counter decreases when the firewall is completely open.
In my experience it is expected at the VM object... it appears the counter was updated to factor in dfw drop packets.
There is no workaround, it appears that is how the metric is built now.
vMan
did you test this? was it the correct answer?
Hi vMan,
sorry for delay response actually I upgraded my vROp's to latest version but still I am seeing some packet droops at VM level
I did little more investigation here by testing vNIC of VM and did not found anything bad but as of now user is not experiencing anything, will update more on it based on my investigation results.
np
did you try to add the VM to the exclusion list in NSX manager and see if the packets drops went away or lowered?
Hi vMan,
I have to get approvals for doing this but will keep updating on it
Thank you,
Pere
ok cool, I was just curious as this was the case of dropping packets for me.
Yeah, as our end users are not seeing any issues, my manager told me to give low priority for this task, may be it was same as you faced.
Thanks for responding for my request
Pere.
cool, I suspect its likely the cause of the dropped packets.
np
you can check the host to verify, if you continue to see dropped packets you might want to google ring buffers and vmware
To check the host
esxcli network vm port list -w "world id"
This will give you the port id for the vm
esxcli network port stats get -p "Port id"
It will return along with other values packets dropped
Packet statistics for port
Packets received: 2445693006
Packets sent: 1635925276
Bytes received: 1388236494062
Bytes sent: 353420219234
Broadcast packets received: 42828139
Broadcast packets sent: 3489
Multicast packets received: 3816956
Multicast packets sent: 128924
Unicast packets received: 2399047911
Unicast packets sent: 1635792863
Receive packets dropped: 7385
Transmit packets dropped: 0