Letbox | Psychz Networks Los Angeles and IPv6 routing trouble(s)

Miky · January 2019

Hello,
Would love to exchange a few words with happy or unhappy users of VPS at the Psychz DC in Los Angeles.

While I have a nice working VPS there, the IPv6 routing seems pretty bad (and I am polite). I'd love to share experience.

Thanks

jure12 · January 2019

I use VPS of 'Letbox' about two years (Psychz DC LA) and everything works properly.
Letbox he has very good VPS (Psychz DC). Letbox has very good network !

Miky · January 2019

@jure12 I'm also quite happy with the VPS from Letbox by itself, just can't succeed to have a working IPv6

hkmix · January 2019

have you tried to submit ticket to them ?

eol · January 2019

@jure12 said:
I use VPS of 'Letbox' about two years (Psychz DC LA) and everything works properly.
Letbox he has very good VPS (Psychz DC). Letbox has very good network !

Sounds like it's optimized.

dahartigan · January 2019

@hkmix said:
have you tried to submit ticket to them ?

I bet my signature the answer to that is a big old no.

Miky · January 2019

@dahartigan said:

@hkmix said:
have you tried to submit ticket to them ?

I bet my signature the answer to that is a big old no.

if you had one, you should start to eat it now!
Of course, there's a ticket, a month and half old. But so far, no change!

dahartigan · January 2019

@Miky said:

@dahartigan said:

@hkmix said:
have you tried to submit ticket to them ?

I bet my signature the answer to that is a big old no.

if you had one, you should start to eat it now!
Of course, there's a ticket, a month and half old. But so far, no change!

Haha, I'm eating my Vegemite sandwich instead (depending on where in the world you're from, that might be considered punishment enough)

Sorry for calling you out for not opening a ticket, I thought it would have been simple probability - virtually all the threads here that start like this have no ticket (or a 19 minute old ticket)

Let it be on the record that I was wrong, take a screenshot, I could edit that admission out later

Miky · January 2019

I like someone able to admit a wrong/too quick judgment... And you are probably and statistically right somehow, newbies do that a lot.

But I'm a dinosaur :-)
And I'm only looking for other users in the same set of IP (v6) and the same kind of trouble.

uptime · January 2019

So @Miky, just to be sure I'm understanding you correctly ...

I'm only looking for other users in the same set of IP (v6) and the same kind of trouble.

You came to LET looking for trouble. And you found @dahartigan. Well played sir, well played indeed.

EDIT2:

Sorry I have nothing useful to add about ipv6 routing on letbox Psychz ... but I'm interested to know more details, and am more than willing to bump your thread with more-or-less pleasant nonsense. Do we have a deal?

EDIT3:

Ah ... on second thought, while you're waiting for fellow letbox ipv6 mavens to give more useful input - in case it might be helpful to compare ipv6 routes with another provider at Psychz - I just setup ipv6 on a box there from CrownCloud, so could run a few traceroutes or whatever if you have some specific locations in mind.

dahartigan · January 2019

@uptime said:
So @Miky, just to be sure I'm understanding you correctly ...

I'm only looking for other users in the same set of IP (v6) and the same kind of trouble.

You came to LET looking for trouble. And you found @dahartigan. Well played sir, well played indeed.

I'll pay that, this time

Miky · January 2019

With a big help from @uptime, I now know that 2 VPS in the same DC "Pschyz Los Angeles" but from 2 different providers (so different IP sets) cannot "talk" to each other via IPv6.

I'm hoping that @Key900 with Letbox, my own provider, and @SpeedBus with CrownCloud, both in that DC, and maybe some more... would help and give me 2 minutes to help diagnose that weird IPv6 routing :-)
Today exemple, From Home/Thailand:

  2    24 ms    22 ms    17 ms  2405:9800:b003:16::18
  3    16 ms    19 ms    17 ms  eth8.vpi.sila1-bgw01.myaisfibre.com [2405:9800:b003:16::17]
  4    18 ms    16 ms    16 ms  aex.iig.sila1-pe01.myaisfibre.com [2405:9800:b003:64::65]
  5    48 ms    46 ms    46 ms  2405:9800:9800:100::7b
  6    44 ms    43 ms    44 ms  6939.sgw.equinix.com [2001:de8:4::6939:1]
  7    77 ms    77 ms    76 ms  100ge14-2.core1.hkg1.he.net [2001:470:0:2ea::1]
  8   117 ms   117 ms   117 ms  100ge10-1.core1.tyo1.he.net [2001:470:0:3c0::2]
  9   227 ms   226 ms   226 ms  100ge11-1.core1.sea1.he.net [2001:470:0:268::1]
 10   225 ms   222 ms     *     comcast-7922-as7922.10gigabitethernet2-12.core1.sea1.he.net [2001:470:0:3c5::2]
 11     *        *        *     Request timed out.
 12     *        *        *     Request timed out.
 13     *        *        *     Request timed out.
 14   245 ms   243 ms   244 ms  be-11599-pe01.losangeles.ca.ibone.comcast.net [2001:558:0:f530::2]
 15   232 ms   231 ms   231 ms  2001:559::a52
 16   233 ms   234 ms   233 ms  gw.2406.8900.lax.us.glonetwork.com.au [2604:6600:0:4::1]
 17   233 ms   232 ms     *     2a0b:ae41::3c
 18   232 ms   232 ms   233 ms  2a0b:ae41::3c

(Just to be clear and fair, I have no problem with Letbox and I am not looking for any kind of fight)

uptime · January 2019

looks like something is off with ipv6 upstream from both letbox and crowncloud at psychz right now.

cause · January 2019

I got /30 subnet allocated for SLAAC and pretty bad egress packetloss only on IPv6. Now i am using he tunnel.
posted here https://www.lowendtalk.com/discussion/154450/ipv6-network-small-reviews

psb777 · January 2019

@Miky said:

... would help and give me 2 minutes to help diagnose that weird IPv6 routing :-)
Today exemple, From Home/Thailand:

  2    24 ms    22 ms    17 ms  2405:9800:b003:16::18
  3    16 ms    19 ms    17 ms  eth8.vpi.sila1-bgw01.myaisfibre.com [2405:9800:b003:16::17]
  4    18 ms    16 ms    16 ms  aex.iig.sila1-pe01.myaisfibre.com [2405:9800:b003:64::65]
  5    48 ms    46 ms    46 ms  2405:9800:9800:100::7b
  6    44 ms    43 ms    44 ms  6939.sgw.equinix.com [2001:de8:4::6939:1]
  7    77 ms    77 ms    76 ms  100ge14-2.core1.hkg1.he.net [2001:470:0:2ea::1]
  8   117 ms   117 ms   117 ms  100ge10-1.core1.tyo1.he.net [2001:470:0:3c0::2]
  9   227 ms   226 ms   226 ms  100ge11-1.core1.sea1.he.net [2001:470:0:268::1]
 10   225 ms   222 ms     *     comcast-7922-as7922.10gigabitethernet2-12.core1.sea1.he.net [2001:470:0:3c5::2]
 11     *        *        *     Request timed out.
 12     *        *        *     Request timed out.
 13     *        *        *     Request timed out.
 14   245 ms   243 ms   244 ms  be-11599-pe01.losangeles.ca.ibone.comcast.net [2001:558:0:f530::2]
 15   232 ms   231 ms   231 ms  2001:559::a52
 16   233 ms   234 ms   233 ms  gw.2406.8900.lax.us.glonetwork.com.au [2604:6600:0:4::1]
 17   233 ms   232 ms     *     2a0b:ae41::3c
 18   232 ms   232 ms   233 ms  2a0b:ae41::3c

So, what seems to be the problem? Isn't that normal?

uptime · January 2019

uptime · January 2019

crowncloud support requested some MTRs to pass along to Psychz ...

Here's a sample from a few minutes ago:

ping6.online.net
Start: Tue Jan 15 09:53:41 2019
HOST: vps.server.com              Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2604:6600:xxxx::1          0.0%    10    0.4   0.4   0.3   0.5   0.0
  2.|-- 2604:6600:0:46::1          0.0%    10    0.8  28.0   0.7 132.8  47.5
  3.|-- 2604:6600:ffff::7         30.0%    10    5.9   6.0   4.0   7.2   1.0
  4.|-- las-b3-link.telia.net      0.0%    10    4.6   4.5   4.0   4.7   0.0
  5.|-- las-b21-v6.telia.net       0.0%    10    7.1   6.5   4.6   7.5   0.7
  6.|-- level3-ic-333036-las-b21. 10.0%    10    7.4   6.5   3.4   7.4   1.2
  7.|-- lo-0-v6.ear2.Paris1.Level  0.0%    10  150.1 149.9 147.6 151.3   0.9
  8.|-- ???                       100.0    10    0.0   0.0   0.0   0.0   0.0
  9.|-- 2001:bc8:0:1::b2          20.0%    10  142.1 143.5 139.6 148.9   3.1
 10.|-- ping6.online.net           0.0%    10  139.7 139.4 138.5 139.9   0.0

.

proof.ovh.ca
Start: Tue Jan 15 09:54:25 2019
HOST: vps.server.com              Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2604:6600:xxxx::1          0.0%    10    0.3   0.4   0.3   0.6   0.0
  2.|-- 2604:6600:0:46::1          0.0%    10    0.7  37.1   0.7 125.1  54.2
  3.|-- 2604:6600:ffff::7          0.0%    10    0.8   2.4   0.8   3.3   0.6
  4.|-- ???                       100.0    10    0.0   0.0   0.0   0.0   0.0
  5.|-- be100-1367.ash-1-a9.va.us 60.0%    10   63.4  59.7  58.1  63.4   2.4
  6.|-- be100-1007.nwk-5-a9.nj.us 90.0%    10   61.2  61.2  61.2  61.2   0.0
  7.|-- be100-1323.bhs-g2-nc5.qc. 10.0%    10   82.1  74.3  72.3  82.1   3.1
  8.|-- vl100.bhs-d2-a75.qc.ca     0.0%    10   71.4  71.1  70.0  71.8   0.3
  9.|-- 2001:41d0:0:50::6:833      0.0%    10   71.2  72.3  69.8  83.5   3.9
 10.|-- 2001:41d0:0:50::2:4049    10.0%    10   71.7  71.8  70.1  75.5   1.4
 11.|-- 2607:5300:60:44a5::1       0.0%    10   71.7  72.8  69.8  88.0   5.3

Miky · January 2019

I have the same ones

ping6.online.net
Start: Tue Jan 15 07:11:10 2019
HOST: letbox                      Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2a0b:ae40::1              10.0%    10    0.6   0.6   0.5   1.3   0.0
  2.|-- ???                       100.0    10    0.0   0.0   0.0   0.0   0.0
  3.|-- las-b3-link.telia.net     20.0%    10    0.6   4.0   0.6  21.1   7.0
  4.|-- las-b21-v6.telia.net       0.0%    10    0.8   1.0   0.7   1.9   0.0
  5.|-- level3-ic-333036-las-b21.  0.0%    10    0.9   0.9   0.9   1.2   0.0
  6.|-- lo-0-v6.ear2.Paris1.Level  0.0%    10  137.6 136.6 136.4 137.6   0.3
  7.|-- 2001:1900:5:2:2::4a0a      0.0%    10  137.5 137.4 137.2 137.7   0.0
  8.|-- 2001:bc8:0:1::b2           0.0%    10  137.2 139.2 137.2 156.8   6.1
  9.|-- ping6.online.net           0.0%    10  137.1 137.0 137.0 137.1   0.0



proof.ovh.ca
Start: Tue Jan 15 07:12:06 2019
HOST: letbox                      Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2a0b:ae40::1               0.0%    10    0.5   2.2   0.5  15.3   4.6
  2.|-- ???                       100.0    10    0.0   0.0   0.0   0.0   0.0
  3.|-- 2001:504:13::210:214      20.0%    10   55.2  55.4  55.1  55.6   0.0
  4.|-- be100-1367.ash-1-a9.va.us 20.0%    10   55.5  55.6  55.3  57.0   0.4
  5.|-- be100-1007.nwk-5-a9.nj.us 90.0%    10   61.1  61.1  61.1  61.1   0.0
  6.|-- be100-1323.bhs-g2-nc5.qc. 10.0%    10   71.6  71.4  71.1  71.9   0.0
  7.|-- vl100.bhs-d2-a75.qc.ca     0.0%    10   70.2  70.2  70.1  70.5   0.0
  8.|-- 2001:41d0:0:50::6:833     20.0%    10   70.0  70.3  69.9  71.7   0.4
  9.|-- 2001:41d0:0:50::2:4049     0.0%    10   70.3  70.6  70.3  71.3   0.0
 10.|-- 2607:5300:60:44a5::1      20.0%    10   70.1  70.1  70.0  70.5   0.0

But I believe the problem is TO Pshyz not FROM.

uptime · January 2019

all the things I don't know about IPv6 could fill a few books and at least one LET thread ... I guess this is going to be the one.

EDIT2:

I just ran a couple inbound MTR to CrownCloud at Psychz (one from Fremont, one from Las Vegas). Seems fine to me but I could be missing something.

Start: Tue Jan 15 07:31:47 2019
HOST: las vegas                    Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2605:6400:xxxxxx::1        0.0%    10    0.7   1.1   0.4   4.7   1.1
  2.|-- 10gigabitethernet5.switch  0.0%    10    1.4   2.1   1.4   5.3   0.9
  3.|-- 10ge1-14.core1.lax2.he.ne  0.0%    10    7.8   8.7   7.8  10.7   0.7
  4.|-- profuse-solutions-llc.gig  0.0%    10   10.4  12.3   7.9  43.6  11.0
  5.|-- lax-us.syn.net.uk         10.0%    10   99.0  24.5   9.0  99.0  30.3
  6.|-- 2604:6600:0:46::2          0.0%    10   12.4  13.2  11.3  17.2   1.5
  7.|-- 2604:6600:xxxx:xxxx::xxx   0.0%    10   21.8  24.2  12.9  32.1   6.4

.

Start: Tue Jan 15 07:29:16 2019
HOST: fremont                      Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- 2602:fed2:xxxx::1          0.0%    10    0.5   0.6   0.4   1.2   0.0
  2.|-- 2001:504:91::9             0.0%    10    0.5   2.6   0.5  19.1   5.8
  3.|-- 100ge9-1.core1.pao1.he.ne  0.0%    10    1.3   1.4   1.1   1.9   0.0
  4.|-- 100ge7-1.core1.lax2.he.ne  0.0%    10   10.2   9.1   8.8  10.2   0.3
  5.|-- profuse-solutions-llc.gig  0.0%    10    9.1  12.3   8.9  37.4   8.8
  6.|-- 2602:ff96:5::1             0.0%    10   10.6  20.7   9.9 102.9  28.9
  7.|-- 2604:6600:0:46::2         10.0%    10   15.4  14.8  13.0  20.6   2.3
  8.|-- 2604:6600:xxxx:xxxx::xxx   0.0%    10   15.1  14.3  13.0  15.6   0.7

jure12 · January 2019

Look at this pretty speed (VPS - letbox - LA DC)

iHavenoName · January 2019

Anyone know when Letbox is going to get more LA vps? Seems to never be in stock.

jure12 · January 2019

@iHavenoName said:
Anyone know when Letbox is going to get more LA vps? Seems to never be in stock.

You can ask here:
https://my.letbox.com/submitticket.php?step=2&deptid=1

EDIT:
These people really have a great network, still in LA or Dallas.

uptime · January 2019

@jure12 - looks decent, I'm assuming reflecting IPv4 network performance. (And nice disk i/o for sure). I hear good things about Letbox but have yet to try them out myself.

This thread really is more intended to be about troubleshooting IPv6 routing for multiple providers at Psychz in Los Angeles.

just getting a few minutes to check again a bit more now - I've got a few updates from support at CrownCloud, investigating IPv6 routing from GT vs Telia but nothing definitive yet.

Miky · January 2019

New morning for me here, and after the first coffee the problem is still around. Despite @uptime 's help, no one succeed to explain why I can IPv6 download to "Letbox/Pschyz/LA" as fast as IPv4, but IPv6 download from "Letbox/Pschyz/LA" is slow as hell!

@jure12 Your screenshot is IPv4 and the working way, sorry!

@uptime Yes, I also - supposedly - got update from Pschyz but so far, it did not change anything for me.

uptime · January 2019

I saw improvement on 3 of 4 download tests (mirror.nl.leaseweb.net, ping6.online.net, proof.ovh.ca all pulling files at least 3 to 5 MB/s to CrownCloud/Psychz via IPv6).

But connection to speedtest.dal01.softlayer.com got stuck somewhere (MTR showing drop at the end of the line at softlayer). At the same time, I was able to pull that testfile via IPv6 at over 5 MB/s to another provider (freerangecloud) in Fremont.

I imagine this sort of whack-a-mole routing troubleshooting is familiar territory for people paying attention to IPv6. I'm wondering what other tools and techniques could be more helpful to pinpoint the glitch(es) ...

Miky · January 2019

uptime said: I imagine this sort of whack-a-mole routing troubleshooting is familiar territory for people paying attention to IPv6. Wonering what other tools might be helpful to pinpoint the glitch(es)?

Yes :-) In a world where IPv6 should be at some professional level, I feel powerless, helpless and kinda "tool-less"!

uptime · January 2019

By the time we figure out IPv6, we'll be saying "I could have had an IPv8"

Letzien · January 2019

@uptime said:
By the time we figure out IPv6, we'll be saying "I could have had an IPv8"

By the time CC supports it, we'll be dead.

uptime · January 2019

https://www.theregister.co.uk/2018/08/28/ipv6_peering_squabbles/

The report specifically notes a long-running dispute between ISP Cogent and Hurricane Electric, as well as another between Deutsche Telekom and Verizon US.

I suspect this doesn't make things any easier for troubleshooting IPv6 routing.

EDIT2:

I know this is old news for people paying attention (definitely not me), but ... is the cogent/HE peering schizm really still dragging on 10 years later? Really?

https://adminhacks.com/broken-IPv6.html

Miky · January 2019

Another morning, another coffee, but same trouble! Exactly 1 month and 5 days I opened that ticket to reach my own provider.

I know I'm bumping my own thread, but if anyone could shed any light he/she would be very much welcome.

Miky · January 2019

@uptime Some news for you, cuz I can't let get so paranoid :-) Cogent and HE may be at war, but it has nothing to do with this thread.
It seems Pschyz are far from the perfect DC, or more exactly they are but only for themselves.

Howdy, Stranger!

Categories

In this Discussion

Letbox | Psychz Networks Los Angeles and IPv6 routing trouble(s)

Comments

Howdy, Stranger!

Quick Links

Categories

In this Discussion

Letbox | Psychz Networks Los Angeles and IPv6 routing trouble(s)

Comments