Opened 2 years ago

Closed 14 months ago

#19760 closed defect (implemented)

Update longclaw's hard-coded IPv6 address

Reported by: teor Owned by: teor
Priority: Medium Milestone: Tor: 0.3.2.x-final
Component: Core Tor/Tor Version:
Severity: Normal Keywords: dir-auth, ipv6, 028-backport, 029-backport, 030-backport, 031-backport, 032-backport
Cc: micah Actual Points: 0.2
Parent ID: Points: 0.1
Reviewer: Sponsor: SponsorZ

Description

In 0.2.8.1-alpha, we added longclaw's IPv6 address as
ipv6=[2620:13:4000:8000:60:f3ff:fea1:7cff]:443
https://gitweb.torproject.org/tor.git/tree/src/or/config.c#n931

Now its descriptor says:
[2620:13:4000:8000:a800:ff:fef5:2213]:443
https://atlas.torproject.org/#details/74A910646BCEEFBCD2E874FC1DC997430F968145

Should we update this before the 0.2.8 release?

Child Tickets

Change History (18)

comment:1 Changed 2 years ago by atagar

For what it's worth this is something we monitor for and Micah has been getting notices...

https://lists.torproject.org/pipermail/tor-consensus-health/2016-July/007288.html

comment:2 Changed 2 years ago by micah

FWIW - I put the old IPv6 address back on longclaw so that this would not be a problem. I wasn't aware that this was being hard-coded into the source, and thought it was fine that it was a temporary address.

I would like to change it to a more permanent address at some point, but right now it is better for me to put the old one back until I have cycles to handle that properly.

comment:3 Changed 2 years ago by nickm

Keywords: TorCoreTeam201608 added; TorCoreTeam201607 removed

No further code or documentation will be written in July, due to time itself. (Leaving needs_revision tickets as-is)

comment:4 Changed 2 years ago by nickm

Keywords: TorCoreTeam201609 added; TorCoreTeam201608 removed

Move unassigned items in August to September.

comment:5 Changed 2 years ago by teor

Keywords: must-fix-before-028 TorCoreTeam201609 removed

We can do this whenever micah is ready, and it will get backported if needed.

comment:6 Changed 2 years ago by teor

Milestone: Tor: 0.2.???Tor: 0.3.???

Milestone renamed

comment:7 Changed 2 years ago by nickm

Keywords: tor-03-unspecified-201612 added
Milestone: Tor: 0.3.???Tor: unspecified

Finally admitting that 0.3.??? was a euphemism for Tor: unspecified all along.

comment:8 Changed 22 months ago by dgoulet

Milestone: Tor: unspecifiedTor: 0.3.1.x-final
Status: newneeds_information

Once micah figure out the IPv6 address that should be used, he'll add it to this ticket. Putting that one in 031 so we don't forget about it!

comment:9 Changed 21 months ago by nickm

Sponsor: SponsorZ

comment:10 Changed 21 months ago by nickm

Milestone: Tor: 0.3.1.x-finalTor: 0.3.2.x-final
Owner: set to micah
Status: needs_informationassigned

comment:11 Changed 21 months ago by nickm

micah, please reassign this to somebody else once there's an IPv6 address we should be using. Thanks!

comment:12 Changed 18 months ago by nickm

Keywords: tor-03-unspecified-201612 removed

Remove an old triaging keyword.

comment:13 Changed 14 months ago by micah

I think that the IPv6 address for longclaw should be removed in the source code. It was only a temporary address that I setup on teor's request for testing and never was meant to be something that stuck. Additionally, longclaw will be moving to a new location, so its ipv6 address will go away then (I'm unsure when it will get a new one), as well as its ipv4 address.

comment:14 Changed 14 months ago by micah

Owner: micah deleted

comment:15 in reply to:  13 Changed 14 months ago by teor

Actual Points: 0.2
Keywords: 028-backport 029-backport 030-backport 031-backport 032-backport added; 029-proposed removed
Owner: set to teor

comment:16 Changed 14 months ago by teor

Status: assignedneeds_review

Replying to micah:

I think that the IPv6 address for longclaw should be removed in the source code. It was only a temporary address that I setup on teor's request for testing and never was meant to be something that stuck. Additionally, longclaw will be moving to a new location, so its ipv6 address will go away then (I'm unsure when it will get a new one)

See my branch longclaw-ipv6-028, which implements this change. The address was added in 0.2.8, so it only needs to be backported that far. We don't need to worry about clients here: 0.2.8 and later have plenty of fallback directory mirrors on IPv6, which isn't on by default anyway.

as well as its ipv4 address.

Now, that will be more interesting. I assume you mean both the hard-coded IPv4 address, and the one in the consensus? Have they changed at all since longclaw was added in 0.2.6.2-alpha / 0.2.5.11 / 0.2.4.26 ?

Here's what clients will do:

  • Clients with a consensus will use whatever address longclaw signs in its relay descriptor
  • Clients without a consensus:

comment:17 Changed 14 months ago by teor

I made a ticket for the IPv4 change at #23592.

comment:18 Changed 14 months ago by nickm

Resolution: implemented
Status: needs_reviewclosed

Merged longclaw-ipv6-028!

Note: See TracTickets for help on using tickets.