Yeah, it is not useful to have an instance with such a high failure rate. I 
need our bots to be announcing failures over IRC or email, and we can’t do that 
with a server that fails daily. It was also losing lots of data points by 
failing.

I agree that we should use a common server.  I will migrate the green dragon 
performance data to a communal server when we have a reliable instance.

I’ll see if I can find a more suitable solution after the break.

> On Nov 25, 2014, at 11:18 PM, Tobias Grosser <[email protected]> wrote:
> 
> On 26.11.2014 00:30, Chris Matthews wrote:
>> I don’t have access to that server.
>> 
>> I have been using a cloud LNT instance for the performance runs on the 
>> green-dragon cluster.  So far it has provided 100% uptime!
>> 
>> http://llvm-lnt.herokuapp.com <http://llvm-lnt.herokuapp.com/>
> 
> I am surprised. What was the reason you did not use the instance at 
> llvm.org/perf? Did you consider it broken?
> 
> If the only issue was that you did not have server access, then maybe we 
> should fix this. I do not see why we would maintain two instances.
> One canonical address for LLVM performance results seems to be the way to go.
> 
> Cheers,
> Tobias
> 


_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to