Open Stack

On 2015-08-31 17:14, George Mihaiescu wrote:
> Is it just dashboard that's slow?
> How about "nova list" or "neutron port-list" from your dashboard node,
> as well as from outside your environment?
> 
> Depending on how you have your endpoints configured in keystone (ip or
> name) and how dns resolution is set in your environment, there might
> be delays in the api calls...

Thanks for the sanity check, George.  You know -- I've gotten so used to 
thinking in terms of performance and GRE tunnels and other abstractions 
that sometimes I just forget to look at the physical layer... where, 
yes, the dashboard node was dropping ~20% of its packets.  Bad port or 
cable (haven't checked which, yet).

Wups.

Don't understand why I didn't see massive retransmits in the pcap, but I 
guess I'm not going to worry myself with that too much.

Thanks, all,

-Ken


> George
> 
> On 31 Aug 2015 15:03, "Ken D'Ambrosio" <ken at jots.org> wrote:
>> 
>> Hey, all.  I've got an Openstack install, with each control node on
> its own 24 GB, quad-core system.  And yet my Horizon is really quite
> slow; I've seen it take 15+ seconds to log in, click on the "routers"
> tab, etc.
>> 
>> The only host among the control nodes that showed significant pain
> was the MySQL host; I replaced its SATA disk with an SSD last night,
> and now instead of a load average ~1.65, it's now around 0.10.  And
> things feel marginally better, but it's still darn slow, so I'm
> guessing the primary bottleneck was *not* MySQL.
>> 
>> Any suggestions on what I should be digging into?  Logs in the cloud
> controller box?  Known API bottlenecks?
>> 
>> Thanks!
>> 
>> -Ken
>> 
>> _______________________________________________
>> Mailing list:
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack [1]
>> Post to     : openstack at lists.openstack.org
>> Unsubscribe :
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack [1]
> 
> 
> Links:
> ------
> [1] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack