Re: [Xen-API] non-persistent stunnel connections

[Dave Scott <Dave.Scott@xxxxxxxxxxxxx>]

[cc:d Rok]

> > Is that happening a lot?
> 
> yes, it for each connection between that slave and the master, multiple
> times per second (so a lot of spams in the slave logs). The are two
> kinds of failures:
> 
> * check_reusable: caught exception
> Http_client.Http_request_rejected("{ frame = false; method = POST; uri
> = /; query = [  ]; content_length = [  ]; transfer encod
> ing = ; version = 1.1; cookie = [  ]; task = ; subtask_of = ; content-
> type = ; host = ; user_agent = xen-api-libs/1.0 }"); assuming not
> reusable
> * check_reusable: caught exception Unix.Unix_error(31, "write", "");
> assuming not reusable

That's very suspicious. I was just talking to Rok about a very similar 
phenomenon in his rrd daemon.

I suspect that the stunnel connection cache is completely ineffective because 
the connections are always timed-out or closed. I think in the long term we 
should get rid of all this stuff and switch to something more transparent, 
perhaps by establishing permanent port-forwards when a host joins a pool. Or 
perhaps we could do something interesting where we create a management VPN 
using the openvswitch somehow.

In the short term we should either disable this buggy mechanism or figure out 
how to fix it...

Cheers,
Dave

_______________________________________________
Xen-api mailing list
Xen-api@xxxxxxxxxxxxx
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api