Wiki uptime is <80%, status page shows no downtime/issues

SatNOGS
1

It seems that the SatNOGS wiki, an extremely important piece of the SatNOGS intrastructure, has very bad uptime. Almost half the times I’ve tried to go to it in the past month, I’ve been unable to, and have ended up having to refer to archive.org versions of it.

Furthermore, it seems that the issue isn’t tracked on the Libre Space status page: https://status.libre.space.

The issue seems like it might be some sort of an overzealous bot filter. If that’s the case though, it’s currently blocking the entirety of my university’s internet, as well as common VPN traffic (and thus should be repaired). Some sort of a more normal CloudFlare bot filter would be far better, where you can complete the captcha if there’s any uncertainty about the user’s authenticity.

curl -4 -v https://wiki.satnogs.org/Main_Page
...
...
curl: (35) TLS connect error: error:0A000126:SSL routines::unexpected eof while reading

Edit: Looks like the issue is finally being tracked on the status page as of today. It wasn’t for the past week though.

2 Likes
2

Hello,

We investigated this and we don’t have anything in place that should block your access.

Are you still having these issues? Can you please provide any more details?

3

It seems that it’s the University of Calgary’s network being sporadically blocked, as connections via VPNs and Tor both make it through fine. AS33091 University of Calgary AS details | IPinfo

The issues are intermittent, and seem very strongly correlated with the university’s network.

What other details would be helpful?

1 Like
4

Hello and thanks for reporting this.

There is no automatic blocking of IPs on the wiki or at least one that we control or are aware of.

Let’s rule out any routing issues first. If you happen to notice such a block again, try to tracepath so we can see that the packets at least reach our host. Last two hops before our wiki should be grnet.gr routers.

I performed the reverse tracepath towards ucalgary.ca which is advertised by the University of Calgary AS. The path is going through GÉANT and there are quite a lot of hops (21!) to reach your university from our side, some of which are asymmetric.

5 
> tracepath wiki.satnogs.org
 1?: [LOCALHOST]                      pmtu 1500
 1:  _gateway                                              1.988ms asymm  2 
 1:  _gateway                                              1.997ms asymm  2 
 2:  no reply
 3:  10.18.x.x                                             1.719ms 
 4:  10.18.x.x                                             0.977ms 
 5:  h1-182.wlan.ucalgary.ca                               1.798ms asymm  6 
 6:  10.x.x.x                                              1.568ms 
 7:  h84.gpvpn.ucalgary.ca                                 1.836ms 
 8:  h97.gpvpn.ucalgary.ca                                 2.080ms 
 9:  no reply
10:  172.69.156.119                                        2.381ms 
11:  172.69.156.120                                        2.029ms 
12:  172.69.156.76                                         2.384ms 
13:  172.69.156.121                                        2.243ms pmtu 1476
13:  no reply
14:  no reply
15:  gnt7-1116.yp3.grnet.gr                              182.356ms asymm 22 
16:  vm2.libre.space                                     187.988ms !H
     Resume: pmtu 1476

The issue also seems like a TLS termination issue (“connection closed early”-type error).

It failed in a browser and with curl before running the tracepath, and works afterwards. Crazy stuff going on.

Thanks for helping take a look at this.

1 Like