CSR 1000v latency issues

[Seittit]

Good morning.

Has anyone else ran into issues with latency on the CSR1000v platforms? Like, REALLY REALLY bad latency?

I have emulated INE's topology with 20 CSR1000v routers within an ESXi hypervisor environment, but packets take several seconds to reach each other. I've added an out-of-band management IP address on an isolated network, but still having the issues. Dr. Google hasn't revealed anyone else experiencing this issues.

Here's router two pinging its peer router 1. These guys are on the same OOB management

Code Select Expand

R2#ping 192.168.1.1
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 192.168.1.1, timeout is 2 seconds:
.....
Success rate is 0 percent (0/5)

bumping up the timeout shows how awful the latency is:

Code Select Expand

R2#ping 192.168.1.1 timeout 2600
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 192.168.1.1, timeout is 2600 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 2390/5905/9766 ms
R2#

[Seittit]

attached: network setup and system resources

it doesn't seem to be an issue with my esx machine nor network utilization.

[Seittit]

amazing, a reboot of the routers solved everything.

Code Select Expand

R2#ping 192.168.1.1
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 192.168.1.1, timeout is 2 seconds:
.!!!!
Success rate is 80 percent (4/5), round-trip min/avg/max = 3/3/4 ms
R2#ping 192.168.1.1
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 192.168.1.1, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 3/3/4 ms

[icecream-guy]

They were sleeping and you woke them up 


remember, if at first you don't succeed, reboot.

[deanwebb]

I need a "have you tried turning it off and on again" smiley. I'd add it now, but I'm about to head out to lunch.

[that1guy15]

Odd, no havent seen that one. But I do reboot about once every few weeks. Also make sure you dont boot too many at once as the they are pretty sensitive about CPU load. To be safe I boot 3 at a time with 5 minutes in between.

[Seittit]

Also make sure you dont boot too many at once as the they are pretty sensitive about CPU load. To be safe I boot 3 at a time with 5 minutes in between.

Good call, I gave my whitebox a mini-heart attack earlier

[that1guy15]

Also make sure you are running the current recommended code. Some of the older code revs are not too stable and have a ton of issues.

Im currently running Version 15.4(3)S1

[wintermute000]

PowerCLI, use a delay interval.... problem solved.

Connect-VIServer -Server 192.168.0.15 -User xxxx\xxxx -Password xxxx
Start-VM -VM SP-LAB-P1
sleep XYZ
!!! repeat

FWIW I think its just a combination of desktop class and non HCL hardware. My CSRs running across 2x sandy bridge i5s and a dvswitch (across a d-link lol, hey it does vlans) average around 30-40ms latency. With Juniper fireflies its up to 70-80ms. I am running off shared storage (iSCSI) as well.

[deanwebb]

Shouldn't this be in the switching and routing section?

[deanwebb]

I need a "have you tried turning it off and on again" smiley. I'd add it now, but I'm about to head out to lunch.

There. It's in the "more" section of smileys. now I'm moving this to R&S.