Investigation of this issue today has now been finished. The root cause of this issue was that a 60G portchannel towards TELIA has been gone down as TELIA had some problems in frankfurt with their equipment. In the same time, our linecard did stop processing of the traffic on that linecard. On the same linecard, also our 100G DECIX port is located. As we had many PFE errors on that FPC, we have restarted this linecard. This did not solve the problem and we contacted the vendors JTAC. They investigated and proposed restarting the whole device to get reinitialized the switching fabric planes, but this did also not solve the issue. A full power cycle was needed, including removing class-of-service configuration, to being able to restart all bgp sessions without getting new PFE errors on this linecard again. So it seems that this is a software related issue, which has to be reviewed by juniper engineers and a software upgrade needs to be performed later on.

The link is working again, nevertheless we are still investigating the root cause of this issue.

The research is still ongoing.

We are currently working with DECIX and the Juniper TAC to see the root cause of this outage. After restarting the linecard facing to DECIX, we see many FPC errors in the log. This seems to be a software relating issue.