Community
Participate
Working Groups
[INFO] Fetching p2.index from https://download.eclipse.org/releases/2019-12/ [INFO] Adding repository https://download.eclipse.org/releases/2019-12 Nov 25, 2019 7:30:17 AM org.apache.http.impl.execchain.RetryExec execute INFO: I/O exception (org.apache.http.NoHttpResponseException) caught when processing request to {s}->https://download.eclipse.org:443: The target server failed to respond Nov 25, 2019 7:30:17 AM org.apache.http.impl.execchain.RetryExec execute
It's not just download.eclipse.org. Access to _all_ *.eclipse.org sites is extremely slow. Gerrit (web & ssh), Bugzilla, Wiki, ... all take several minutes to load, if they load at all.
Seems like this was an intermittent hick up. All services look much better again now. Can you confirm?
Yes, appears to be back. But look at https://status.eclipse.org/ . There was definitely something strange going on from around 07:00 to 10:00. Web page access seems to be OK again, but download speed? Perhaps the graph is not up to date; don't notice any unusual slowness in running CI builds, which also download from downloads.eclipse.org.
(In reply to Thomas Wolf from comment #3) > Yes, appears to be back. But look at https://status.eclipse.org/ . There was > definitely something strange going on from around 07:00 to 10:00. > > Web page access seems to be OK again, but download speed? Perhaps the graph > is not up to date; don't notice any unusual slowness in running CI builds, > which also download from downloads.eclipse.org. I'm not denying that there was an issue. We will look into it.
A DoS attack was triggered at 1:51am localtime (UTC-5), originating from South Korea. We'll be filing an abuse claim to the netblock owners. Timeline of events: 01:51: Eclipse firewall blocks one attacking IP address in South Korea 02:13: Eclipse firewall dropping >27,000 attack packets per second. CPU usage maxes out and legitimate traffic flow is affected 02:30: Attack persists, legitimate traffic flow is severely degraded. Timeouts begin to occur. 04:03: Attack ceases. Traffic flows return to normal. Our Firewall, which is about 10 years old, does not provide adequate protection for this type of DoS (malicious or not; see bug 515596). We are tracking a replacement since 2017-09. For now, worksforme is all I can do.