- ripe-atlas - mailman.ripe.net

Ping Fails (ns.ripe.net)
by AlbyVA 23 Jul '13

23 Jul '13

In reviewing My Probe stats, it appears that (ns.ripe.net) was once able to ping the server, but now it appears to be failing. All my other ping stats under My Probe show successful pings to other servers, but for some reason ns.ripe.net is failing. Does anybody have a suggestion of what might be going on, since ping is working to all the other servers. -Alby

4 6

No address in the results when ping or traceroute fail?
by Stephane Bortzmeyer 23 Jul '13

23 Jul '13

When a ping or traceroute emasurement fails, we do not know the IP address of the probe (empty "from" and no "src_addr"), which is annoying when you want to analyze the possible causes of the failure: { "size": 20, "addr": "2620:0:2d0:200:0:0:0:7", "min": -1, "rcvd": 0, "from": "", "dst_name": "2620:0:2d0:200:0:0:0:7", "af": 6, "timestamp": 1374395985, "fw": 4520, "proto": "ICMP", "name": "www.icann.org", "prb_id": 3953, "result": [ { "error": "sendto failed: Network is unreachable" }, { "error": "sendto failed: Network is unreachable" }, { "error": "sendto failed: Network is unreachable" } ], "max": -1, "dup": 0, "avg": -1, "type": "ping", "dst_addr": "2620:0:2d0:200::7", "sent": 0, "msm_id": 1013576 },

3 6

No probe shown in My Probes list (Registered. Plugged Probe v3 to network. 1st, 3rd, 4th LEDs lighted up.)
by Arthit Suriyawongkul 22 Jul '13

22 Jul '13

Hi I'm Arthit from Bangkok. New here. The switch at the probe is now at "3G/4G". Should i change to WISP or AP? thanks, Art -- เครือข่ายพลเมืองเน็ต -- "เปิดเน็ต เปิดใจ" Thai Netizen Network -- "Open Net. Open Mind." https://www.facebook.com/thainetizen

3 4

Strange and undocumented (?) result in traceroute JSON output
by Stephane Bortzmeyer 17 Jul '13

17 Jul '13

Measurement #1013405 contains: "result": [ { "x": "*" }, { "late": 20, "ttl": 61, "from": "88.79.15.141", "size": 96 }, { "x": "*" }, { "x": "*" } While "late" is documented <https://atlas.ripe.net/doc/data_struct#v4400_traceroute>, the lack of "rtt" field is not and crashes tools like json2traceroute.py.

3 2

[UDM] What value to give to 'requested' when 'type' = 'msm'?
by Stephane Bortzmeyer 17 Jul '13

17 Jul '13

The documentation <https://atlas.ripe.net/doc/measurement-creation-api/> says: "probes": [ { "requested": 1, "type": "msm", "value": 1000002 } ] but, from my tests, if you do so, you get only one probe, whatever measurement #1000002 actually used. To get all the probes, I had to indicate a very large number in 'requested' (unless you remember or count the actual number).

2 2

IPv6 and Atlas probe v3
by Mike. 16 Jul '13

16 Jul '13

My probe works fine on IPv4, obtaining IP address, router, and DNS servers via DCHP. The IPv6 capability has me curious with two questions... I statically configured the IPv6 address via the web interface. I gave the probe an IPv6 address from my prefix, a gateway address, and the IP address of a DNS recursive server provided by my tunnel vendor. The IPv6 address shows up OK on the web interface ("Internet Address") The AS Number also shows up fine. The Gateway is "Undetermined/Unknown" The IP addr of the DNS server is listed as shown below. In the text table below the configuration info, I see: ("V" is a check, "X" is an X) Static Enabled Probe Knows Probe Uses Last Static Address Given IPv6 V V V 2001:470:xxx:xxx::xxx DNS V V X 2001:470:20::2 Question #1: Why is there an "X" for DNS in the "Probe Uses" column? Is there something wrong with that DNS server? When I query DNS from another box on the same subnet as the Atlas probe I get valid results: ; <<>> DiG 9.8.3-P4 <<>> @2001:470:20::2 atlas.ripe.net any ; (1 server found) ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 19739 ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 0 ;; QUESTION SECTION: ;atlas.ripe.net. IN ANY ;; ANSWER SECTION: atlas.ripe.net. 5922 IN AAAA 2a01:4f8:121:30a3::78:16 atlas.ripe.net. 5911 IN A 178.63.78.16 ;; Query time: 21 msec ;; SERVER: 2001:470:20::2#53(2001:470:20::2) ;; WHEN: Mon Jul 8 12:24:01 2013 ;; MSG SIZE rcvd: 76 I also noticed this sentence at the bottom of the https://atlas.ripe.net/doc/static-config page: "...As of now, we have no way of preventing the probe from accepting RAs. So if there is an IPv6 (rogue) RA on you network, it can/will override part of the static configuration. We're unsure if this will change in the future. ..." I do have RA support on the subnet, though I would not classify it as "rogue". :) The RA provides the IPv6 prefix, prefix length, and the IPv6 address of the recursive DNS server. This RA support has been used by another box on the subnet, and it seems to work fine. Question #2: Does the Atlas v3 probe use RA if they are present? Thanks.

2 2

Just to say thank you
by Stephane Bortzmeyer 16 Jul '13

16 Jul '13

With Atlas probes, we were able (at the cost of a few credits) to solve a difficult network connectivity problem. I just wanted to say THANKS.

1 0

Re: [atlas] Timeout on Ping Measurements
by Philip Homburg 15 Jul '13

15 Jul '13

On 2013/07/14 19:35 , Rusty Dekema wrote: > > What is the default timeout on ICMPv4 user-defined measurements made > on RIPE Atlas? I am guessing something like 1.2 s based on > observations, but I'd like to know definitively. It should be 1 second. > Also, is it possible to increase the timeout? One of the targets I > have been measuring often takes a very long time to respond (due to > congestion or link-layer retries) and I would like to have a high (5 > second, perhaps) timeout on these pings. I haven't noticed an option > to set this when setting up the measurement, but it's very possible > I've overlooked something. At the moment there is no option to change the ping timeout. Traceroute does have that option. Philip

1 0

Timeout on Ping Measurements
by Rusty Dekema 14 Jul '13

14 Jul '13

Greetings, What is the default timeout on ICMPv4 user-defined measurements made on RIPE Atlas? I am guessing something like 1.2 s based on observations, but I'd like to know definitively. Also, is it possible to increase the timeout? One of the targets I have been measuring often takes a very long time to respond (due to congestion or link-layer retries) and I would like to have a high (5 second, perhaps) timeout on these pings. I haven't noticed an option to set this when setting up the measurement, but it's very possible I've overlooked something. Cheers, Rusty Dekema

1 0

[iesg-secretary@ietf.org: [lmap] WG Action: Formed Large-Scale Measurement of Broadband Performance (lmap)]
by Stephane Bortzmeyer 11 Jul '13

11 Jul '13

It seems quite close from what Atlas is doing. Atlas documentation in a RFC, one day? :-)

2 1