We expect to have the issue resolved by tomorrow – the network admins found a problem between XO and TInet.
Just an update.
The issue is still ongoing, the buffering happens for a few minutes a few times a day still – if the datacentre don’t have this fixed within 48 hours then the contract will be cancelled in the next month and customer accounts will be moved to a new facility.
I do apologise to customers affected by this, I do feel your frustration
There have been some issues on the usa prem network lately, until today the engineers there suggested that it was a problem with the servers but I’ve managed to get evidence together that shows that it’s the network.
They are currently working to find the cause of this intermittent buffering that we’ve experienced over the last couple of weeks – it only seems to happen for a few minutes at a time, but I’ve told them how irritating this is for people that use the network for streaming.
I expect this to be resolved soon, and if not then customers will be relocated.
Sorry for any inconvenience caused.
I’m not 100% certain but there may be an issue with the USA premium network, a customer has reported issues with AutoDJ and I’ve heard a little buffering on another stream. Thre’s no packet loss and there isn’t any lag but something is amiss.
I’ll contact the network people and keep you updated here.
UPDATE: I have a ticket open with the datacentre, who will look in to this shortly. Please don’t open a support ticket, as it may not be answered right away if it is regarding this issue.
According to Radiotoolbox stream test, the streams are perfect http://www.radiotoolbox.com/images/sbin/stream_test_graph.png?id=13119 — this does suggest that only certain routes are affected.
UPDATE3 22:09: Appears to be fixed
The issue seems to have cleared up at the moment, I’ll post here with any info as I get it.
UPDATE4 22:19:
The issue was with one of the carriers, XO – it seems to have calmed down now and they are looking in to the problem. The problem was outside the control of us or the datacentre as it was on an international link outside the network.
Due to a problem with the server resetting a few times over the last week, we are scheduling an engineer to power down the server and check out the hardware. The scheduled downtime will be:
2:00AM, Tuesday 22nd September 2009.
10:00PM, Monday 21st September 2009.
We expect the downtime to be around 30 minutes, and once it is powered up, all AutoDJ streams on the server will restart automatically.