Good point, that still leaves the job to manage the excess of
connections on the user side but at least it will fail in an ordered and
well known way.
Another possibility I was considering is to retrieve all the instances
for a given deltacloud instance and do the job locally on my side, the
question then is, how does the time it takes to retrieve the "all
instances" blob grows in relation with the number of instances? Right
now it takes some aprox 15 seconds to retrieve aprox 50 instances in
JSON form, is it linear, exponential or does it stays stable?
Xavier
On 10/26/2012 10:16 AM, Michal Fojtik wrote:
On 10/26/2012 10:05 AM, Xavier Naveira wrote:
Hi,
I'm building a web application that uses deltacloud to poll status for
the servers on a vCenter 5.0.0
If I launch too many queries concurrently vCenter stops answering them,
I've googled around and it seems taht vCenter has a max concurrent
connections limit that is 40.
I know that you could argue that this is not a deltacloud issue but if
you are pursuing total transparency for the deltacloud users it would be
a really nice feature that deltacloud itself does the management of this
limit.
What do you think?
I think you're right + this is 'good to known' :-) We can expose the
--max-conn thin option using 'deltacloudd' command, so when you start DC
you can do:
$ deltacloudd -i vsphere --max-conn 40
-- Michal
