noticeboard.ru.ac.za

2010/05/06 - Unscheduled Internet Outage
At about 10AM this morning the University lost access to the Internet for reasons as yet unknown. Users will likely experience this outage as an ERR_DNS_FAIL or ERR_CONNECT_FAIL error when trying to access websites. When accessing secure (https://) web sites, your browser may report that the connection has timed out or the site is unreachable.

We've reported this outage to TENET's service desk.
Neotel have re-routed our connection via an alternative protection circuit, and Internet access has been restored. We remain at risk, however, since the circuit we normally use is still down.
We've just lost access again. We've reported this to TENET's service desk
We're informed that the problem is power-related and that Eskom are en-route to repair. Whilst we have access again at the moment. we should expect another outage whilst Eskom are working. They expect to be on site at 11:45.
At about 3:15PM we lost access for the third time today. We've once again reported the problem.
Considering this is a continuation of the fault we reported at about 10AM this morning, we've just asked TENET to escalate the matter for us and try and get some definitive answers.
We're told that the root cause of our outages is that Eskom is working on power lines between Cape Town & Port Elizabeth. Infraco's fibre optics, which Neotel use to connect us, run along these power lines and are being affected by Eskom's work.

The circuit in question is supposed to be protected (we've been repeatedly assured that it is). Protection would mean that, whilst we'd be at increased risk of an outage, maintenance work such as this would not affect us because we'd automatically re-route via an alternative circuit. The primary circuit from Port Elizabeth to Cape Town runs along the coast; the protection circuit supposedly goes inland via Bloemfontein. Given the number of outages we've had that involve work along the coast (particularly around Blanco/George), clearly this circuit isn't protected or the protection does not work properly.

Neotel have been unable to get an estimated time-to-restore from Infraco.

Neotel have also just passed on a request from Infraco to do scheduled maintenance between 7AM and 11AM this coming Sunday. As this does not fall within the scheduled maintenance window (midnight to 6AM), we've refused their request and asked them to reschedule for a later date (and preferably within the scheduled maintenance window, or at least starting at first light). Past experience, however, shows that they're likely to ignore our contractual right of refusal.

We've asked TENET to explicitly take these issues up with Neotel's service management at the next service review, and that we be included in the service review meeting.
TENET have just agreed to escalate this to Neotel's service manager, the highest operational level of escalation.
We've just been informed that the problematic link is the one between "Louter-water" (Tsitsikama) towards "Melkhout" (Humansdorp). These are both Eskom substations along the N2 coast route, and on the part of our circuit that is supposed to be protected.

Apparently there is an electrical problem that is causing the termination equipment to intermittently fail. There was a stand-by generator on site, but it failed to start and is currently not operational. Neotel/Infraco/Eskom (we're losing track of whose equipment is whose) are currently trying to source an alternative stand-by generator.

It is likely that our link recovers and drops with as electrical power does, and will continue to do so whilst Eskom work on the power line. Whilst we currently have access, expect outages during the course of this evening and until this problem is resolved.
We've just been told that the problem at louter-water has been resolved, and that there's now a functioning stand-by generator on site.
There are now some interesting statistics of the service we've received from Neotel during the GEN3 period available.
post.5532221