OVHcloud Network Status
Current status
Legend
- Operational
- Degraded performance
- Partial Outage
- Major Outage
- Under maintenance
FS#8615 — cdns on RBX1
Scheduled Maintenance
Report for Network & Infrastructure
Completed
During mitigation, attacks on some RBX1 customers
have revealed a routing problem between CDNS
213.186.33.99 and RBX1: the packets leave the
internal routing in order to pass outside the DC,
then re-enter the DC again. Not a serious issue in normal
time, but it's affecting the servers undergoing mitigation
tests, as these packets pass through the mitigation and
mix with the DNS AMP attacks thus filtered.
This problem is linked to the RBX1 routing architecture
which dates back to 2006 and is very flexible. RBX1
routing will be upgraded in June.
RBX2/3/4/5 have not been affected thanks to an
internal network that joins all the DC routers
together, including these sites: BHS, RBX, SBG.
All this traffic remains internal, except RBX1 which
is not configured in the same way as the rest, and
cannot be changed without causing 2hrs downtime.
In the meantime, we have done a fix
by put the dedicated CDNS in place in RBX1.
The server is obviously too weak to
withstand all the RBX1 requests. We are in the process
of upgrading it onto a more powerful server.
have revealed a routing problem between CDNS
213.186.33.99 and RBX1: the packets leave the
internal routing in order to pass outside the DC,
then re-enter the DC again. Not a serious issue in normal
time, but it's affecting the servers undergoing mitigation
tests, as these packets pass through the mitigation and
mix with the DNS AMP attacks thus filtered.
This problem is linked to the RBX1 routing architecture
which dates back to 2006 and is very flexible. RBX1
routing will be upgraded in June.
RBX2/3/4/5 have not been affected thanks to an
internal network that joins all the DC routers
together, including these sites: BHS, RBX, SBG.
All this traffic remains internal, except RBX1 which
is not configured in the same way as the rest, and
cannot be changed without causing 2hrs downtime.
In the meantime, we have done a fix
by put the dedicated CDNS in place in RBX1.
The server is obviously too weak to
withstand all the RBX1 requests. We are in the process
of upgrading it onto a more powerful server.