- ris-users - mailman.ripe.net

[maintenance] RRC22 in Bucharest, 17 April 2024 - 08.00 - 16.00 UTC
by Marco Giuliani 11 Apr '24

11 Apr '24

Dear all, On April 17th, 2024, we will reinstall rrc22.ripe.net, the RIS collector hosted at InterLAN Bucharest. AS12654 86.104.125.61 2001:7f8:64:225::61 Work start: 2024-04-17 08:00 UTC Work end: 2024-04-17 16:00 UTC Kind regards, Marco Giuliani RIPE NCC

1 0

RIS beacon ROA configuration error
by Martin Pels (RIPE NCC) 04 Mar '24

04 Mar '24

Dear all, As documented on [1], we operate several RIS Routing Beacons for RPKI. Due to a configuration error, two of the beacons were advertised with an incorrect RPKI status for several days. This concerns the "RPKI Unknown" beacons, which are the following prefixes: * 2001:7fb:ff03::/48 * 84.205.83.0/24 For both these prefixes, a ROA was created at 2024-02-27 7:55 UTC, turning these announcements into RPKI Valid. Upon receiving a report about this incorrect change we have removed the ROA again. This was done at 2024-03-04 11:31 UTC. Kind regards, Martin Pels RIPE NCC [1] https://ris.ripe.net/docs/routing-beacons/#special-beaconing-setups

1 0

[maintenance] RRC10 in Milan, 21 February 2024 - 08.00 - 16.00 UTC
by Marco Giuliani 19 Feb '24

19 Feb '24

Dear all, On February 21st, 2024, we will replace the hardware of rrc10.ripe.net, hosted at MIX - Milan Internet eXchange. AS12654 217.29.66.6 2001:7f8:b:100:1d1:a5d1:2654:6 Work start: 2024-02-21 08:00 UTC Work end: 2024-02-21 16:00 UTC Kind regards, Marco Giuliani RIPE NCC

1 0

[maintenance] RRC22 in Bucharest - 7 February 2024 - 10.30 - 12.30 UTC
by Marco Giuliani 05 Feb '24

05 Feb '24

Dear all, rrc22.ripe.net, hosted at InterLAN Bucharest, will be down for two hours on February 7th, 2024, to let the local technical staff move the hardware to a different rack. AS12654 86.104.125.61 2001:7f8:64:225::61 Work start: 2024-02-07 10:30 UTC Work end: 2024-02-07 12:30 UTC Kind regards, Marco Giuliani RIPE NCC

1 0

RIS anchor/beacon changes
by Martin Pels (RIPE NCC) 02 Feb '24

02 Feb '24

Dear all, In the coming months we will be making several changes in the prefix advertisements by RIS Route Collectors. These changes were discussed previously on RIPE Labs[0]. January 2nd 2024 ---------------- The following BGP announcements will be disabled: * IPv4 anchor+beacon announcements on RRC04, RRC07, RRC10, RRC13, RRC14, RRC15, RRC16, RRC21 and RRC22 * IPv4 anycast failover simulation beacon (84.205.72.0/24) on RRC03 and RRC14 All IPv6 anchor+beacon announcements will remain. February 1st 2024 ----------------- We will be enabling the following new prefix advertisements: * A single anycast IPv4 and IPv6 anchor+beacon pair per RIR region, advertised from each RRC in this region. * An anycast IPv4 /25 and /28 anchor, advertised from one RRC per RIR region * An anycast IPv6 /56 and /64 anchor, advertised from one RRC per RIR region To accommodate these changes we will also modify the encoding used in the BGP AGGREGATOR attribute. For full details and an overview of all anchor and beacon prefixes, please refer to our updated documentation[1]. Kind regards, Martin Pels RIPE NCC [0] https://labs.ripe.net/author/emileaben/ris-beaconing-shining-a-light-on-bgp… [1] https://ris.ripe.net/docs/routing-beacons

1 1

RIPE RIS MRT download speed issue from US
by Mingwei Zhang 17 Jan '24

17 Jan '24

Dear RIS engineer/users, Greetings! I've been seeing inconsistent and slow download speed when trying to access RIPE RIS MRT files from the US. I'd like to share what I observed and I hope this could be helpful in resolving the underlying problem. In early January, I noticed a consistent sub 500KBps download speed, observed from multiple locations myself (from boxes at ATT, Equinix Metal, and via Cloudflare WARP). I also posted a quick twitter survey and received multiple responses saying the same https://twitter.com/heymingwei/status/1742959348396220557. The situation seems to be resolved a few days later, going back to 20MBps speed at home for me. Today, I again see consistent <1MBps speed from the US west coast at home with 1Gbps symmetric link (AS7018 ATT). [image: image.png] Here is the traceroute from my location (no VPN, 1Gig ATT fiber) to data.ris.ripe.net: ➜ /tmp traceroute -a -q 1 data.ris.ripe.net > traceroute to ftp.ripe.net (193.0.11.24), 64 hops max, 52 byte packets > 1 [AS0] unifi (10.0.0.1) 2.858 ms > 2 [AS0] 192.168.1.254 (192.168.1.254) 0.899 ms > 3 [AS0] 108-238-20-1.lightspeed.sndgca.sbcglobal.net (108.238.20.1) > 4.446 ms > 4 [AS7132] 71.157.6.120 (71.157.6.120) 4.558 ms > 5 [AS7018] 12.123.215.174 (12.123.215.174) 16.459 ms > 6 [AS7018] la2ca22crs.ip.att.net (12.122.2.69) 14.652 ms > 7 [AS7018] ggr2.la2ca.ip.att.net (12.122.129.105) 10.493 ms > 8 [AS0] ae7.cr4-lax2.ip4.gtt.net (199.229.231.29) 9.350 ms > 9 [AS3257] ae1.cr3-ams2.ip4.gtt.net (89.149.128.117) 170.055 ms > 10 * > 11 * > 12 * > 13 * Please let me know if there is any other info that I can provide that may help. Cheers, Mingwei

2 1

RIS Quarterly Planning Q1 2024
by Paul de Weerd 20 Dec '23

20 Dec '23

Dear colleagues, The quarterly plans for RIS can now be found at: https://www.ripe.net/support/documentation/quarterly-planning/ris In this quarter, we'll continue working on modernising the RIS beaconing and open-sourcing the RIS Live code. We hope that by providing these plans, we can better inform the community of the ongoing developments on RIS and get feedback on our work. If you have any input on these plans or suggestions for future work, please reach out to us. Kind regards, Paul de Weerd RIPE NCC

1 0

Re: ris-users Digest, Vol 50, Issue 3
by Rob Ballantyne 29 Nov '23

29 Nov '23

Thanks Colin, I tried today, unsuccessfully unfortunately, to get a small example to fail in the way my stuff was failing. If I can prove out a bug, I’ll definitely create Github issue. Thanks for the help! Rob > > ------------------------------ > > Message: 2 > Date: Tue, 28 Nov 2023 11:43:00 +0100 > From: Colin Petrie <colin(a)spakka.net> > To: Rob Ballantyne <robballantyne3(a)gmail.com> > Cc: ris-users(a)ripe.net > Subject: Re: Restrictions around freeing memory with > bgpdump_free_memory > Message-ID: <2c44746b4c0a70be113a36080fd01539(a)spakka.net> > Content-Type: text/plain; charset=UTF-8; format=flowed > > Hi Rob, > > bgpdump versions before 1.5.0 had thread-safety issues due to use of > global variables. If you are still having issues with the latest > version, please open an issue on Github so we can look into the problem. > > Thanks, > Colin > >

1 0

Restrictions around freeing memory with bgpdump_free_memory
by Rob Ballantyne 28 Nov '23

28 Nov '23

Hi, I’m running into memory access exceptions when using bgpdump_free_memory(). I’m collecting an interval worth of BGPDUMP_ENTRYs (1 minute worth) into buckets. After I have a bucket full I do a pairwise computation (that can take some time) of just this buckets data on a thread (from a pool) while the main thread moves onto the second bucket (and then starts up potentially another thread on the second bucket). … After the thread completes its summary statistics computation, it goes to call bgpdump_free_memory() on each of the BGPDUMP_ENTRY. I mutex lock so that only on thread will execute in bgpdump_free_memory() at a time. Are there restrictions (express or implied) on how to call bgpdump_free_memory() whether multi-threaded or not? BTW, if I run just a single thread it does seem to work just fine. Thanks, Rob

2 1

Two years of selective peering with RIS
by Michela Galante 23 Nov '23

23 Nov '23

Dear colleagues, We published an article on RIPE Labs where we talk about our RIS peering strategy of the last two years, what we've achieved so far, and why we've temporarily removed the peering form from our website: https://labs.ripe.net/author/emileaben/two-years-of-selective-peering-with-… We would love to hear your thoughts about it either here on this mailing list, in the Forum or live at RIPE 87. Best regards, Michela Galante RIPE NCC

1 0