Once upon a time, Max Vernimmen <[email protected]> said:
> You mention that you don't see any issues with the switches. Are you tracking 
> the amount of frames that get dropped by your switches? Any other errors on 
> the switch ports? How's the link utilisation?

Yes, I'm tracking that, and there are no drops at the switches (or the
Linux servers' end).

The link utilization is pretty low.  The server 1G ports are typically
only running about 5-20 Mbps on each link (a spike is 50-100 Mbps on
each, and those are usually only on one server at a time, hosting a
busier VM).  The SAN 10G ports are running about 30-50 Mbps (spiking to
maybe 200 Mbps).  My monitoring system is pulling stats from the switch
every 60 seconds.

I could reproduce some of the latency on a server in maintenance mode,
where my script was the only thing touching the disk (so no other
storage traffic from that server).

> When you get high latency, does it affect only that node or all at the same 
> time? In that case, do you have pause frames on and are they being triggered? 
> Is there a spike in traffic at that moment (that second)?

It seems to affect multiple nodes at the same time, although not always
(or at least not always to the same extent).  Usually, the oVirt
warnings about high latency are only on one node at any given time (but
there has been a time or two where it has been two nodes).

Looking at pause frame stats on the switch, they're essentially zero
(0-100 pause frames received on some server ports with a switch uptime
of 5 months).  No pause frames on the SAN ports, and no pause frames
transmitted on any port from the switch.

> Is there any packetloss between your hosts and your eql(s) ?

Not that I see.

> Those were just some thoughts that popped into my head. I hope they can help 
> you find the cause.

Thanks.  I'm reasonably sure it isn't a network issue, but I hadn't
checked all of the above.
-- 
Chris Adams <[email protected]>

_______________________________________________
Linux-PowerEdge mailing list
[email protected]
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

Reply via email to