View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0004728 | unreal | ircd | public | 2016-08-11 17:05 | 2016-11-10 09:29 |
| Reporter | Gottem | Assigned To | syzop | ||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | unable to duplicate | ||
| Product Version | 4.0.5 | ||||
| Summary | 0004728: Netsplit messes up hostmask | ||||
| Description | After a netsplit, some users' vhosts get funky. It shows properly in (some) whoises, but never in a PRIVMSG or JOIN line. They tend to have ZNC with auto-identify configured. Here's some raw output: :[email protected] QUIT :*.net *.split :[email protected] JOIN :#ayylmao :[email protected] KICK #ayylmao shrink :Test :[email protected] JOIN :#ayylmao But then when I run a whois (nickname twice to query the server they're on directly): :[email protected] WHOIS shrink :shrink :leaf.dom.tld 311 Gottem shrink ple se.fuck.off * :Peter And a "simple" whois through my server: :[email protected] WHOIS shrink :other.dom.tld 311 Gottem shrink ple halp-EB6DA54D.us-west-2.compute.amazonaws.com * :Peter So two servers, two different hostnames. I'm not entirely sure if it's Anope or Unreal related, but it looks like the vhost isn't propagated properly by Unreal. | ||||
| Steps To Reproduce | - Request a vhost from HostServ - Enable auto identify - Make sure you're connected to a leaf, then SQUIT it from the hub - Disconnect and reconnect before the srv2srv link is back up - You'll auto identify when the link comes back and your vhost is applied (partially) | ||||
| 3rd party modules | |||||
|
|
I've trouble reproducing this. With auto-identify you mean /NS IDENTIFY when nickserv says "This nickname is registered and protected." right? I'm connected and identified on a leaf, squited the server, reconnected the client (thus no longer identified), as you said. Then identifying upon the "This nickname is registered and protected." message. My vhost is applied. If I then /WHOIS myself (on the leaf) or /WHOIS mynick (on the hub) the vhost is correctly shown. |
|
|
Yeah, that's what I meant. You may need to open another connection though, as the vhost does show properly in whoises. It's JOIN/PART messages where it gets funky. |
|
|
Still no luck here. I have 3 sessions open on 3 different servers and the host is the same on all if I /WHOIS or if I let the user speak and check the raw traffic (:nick!user@vhost PRIVMSG ..) I also made sure I tested with and without +x set (before receiving the vhost), makes no difference. I use the following linking topology: A `-HUB `-services Also tried: A `-B `-HUB `-services In all cases the user is on A and I squit A's connection to next server. What anope version do you use? Any other tips? |
|
|
I'm having a little trouble manually reproducing it again as well. We did grab some of your IRCd updates, so perhaps you inadvertently caught this bug with it. We'll keep an eye out for any further occurrences. We're on Anope v2.0.3 by the way. |
|
|
Ok, same version as me. |
|
|
As per above: assuming it has been fixed in a previous release, in any case it is no longer reproducible. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2016-08-11 17:05 | Gottem | New Issue | |
| 2016-10-01 11:40 | syzop | Note Added: 0019444 | |
| 2016-10-01 11:40 | syzop | Note Edited: 0019444 | |
| 2016-10-01 11:41 | syzop | Assigned To | => syzop |
| 2016-10-01 11:41 | syzop | Status | new => feedback |
| 2016-10-01 11:42 | syzop | Note Edited: 0019444 | |
| 2016-10-01 11:43 | syzop | Note Edited: 0019444 | |
| 2016-10-01 11:43 | syzop | Note Edited: 0019444 | |
| 2016-10-01 13:33 | Gottem | Note Added: 0019445 | |
| 2016-10-01 16:41 | syzop | Note Added: 0019446 | |
| 2016-10-01 17:12 | Gottem | Note Added: 0019447 | |
| 2016-10-01 18:56 | syzop | Note Added: 0019449 | |
| 2016-11-10 09:29 | syzop | Status | feedback => closed |
| 2016-11-10 09:29 | syzop | Resolution | open => unable to duplicate |
| 2016-11-10 09:29 | syzop | Note Added: 0019500 |