l2tpns issueshttps://code.ffdn.org/l2tpns/l2tpns/-/issues2022-09-02T23:19:11Zhttps://code.ffdn.org/l2tpns/l2tpns/-/issues/11Make sure releases are sane2022-09-02T23:19:11ZzorunMake sure releases are sanel2tpns 2.3.3 in buster-backports advertises itself as 2.3.2 in the CLI:
```
router> show version
L2TPNS 2.3.2
```
Probably the version number was not bumped during the release.
We could maybe script the release process? Or add a CI t...l2tpns 2.3.3 in buster-backports advertises itself as 2.3.2 in the CLI:
```
router> show version
L2TPNS 2.3.2
```
Probably the version number was not bumped during the release.
We could maybe script the release process? Or add a CI task that catches this kind of mistakes?https://code.ffdn.org/l2tpns/l2tpns/-/issues/12`cin_delta` / `cout_delta` should be extended to 64bit2022-07-02T12:16:55Zsthibaul`cin_delta` / `cout_delta` should be extended to 64bitHello,
Our FTTH stats show up as capped around 100Mbps. This is apparently simply coming from `cin_delta` and `cout_delta` being 32bit: with a report every 5 minutes, that's 4GiB/5minutes ~= 100Mbps.
`cin_delta` and `cout_delta` should...Hello,
Our FTTH stats show up as capped around 100Mbps. This is apparently simply coming from `cin_delta` and `cout_delta` being 32bit: with a report every 5 minutes, that's 4GiB/5minutes ~= 100Mbps.
`cin_delta` and `cout_delta` should thus be extended to 64bit integers, to handle accounting with nowadays bandwidths :)
That doesn't seem completely trivial since the `sessiont` structure is shared in clusters and a version number seems to be used to automatically convert, see `cluster.c`
Samuelhttps://code.ffdn.org/l2tpns/l2tpns/-/issues/9Invalid prefix sent to BGP peer when using bind_address2021-10-15T19:55:46ZdomiInvalid prefix sent to BGP peer when using bind_addressIn l2tpns.c (line 4140) :
> if (config->bind_address)
> bgp_add_route(config->bind_address, 0xffffffff);
Will produce invalid BGP update :
> Network Layer Reachability Information (NLRI)
> NLRI length...In l2tpns.c (line 4140) :
> if (config->bind_address)
> bgp_add_route(config->bind_address, 0xffffffff);
Will produce invalid BGP update :
> Network Layer Reachability Information (NLRI)
> NLRI length 255 invalid (> 32)
> [Expert Info (Error/Malformed): NLRI length 255 invalid (> 32)]
> [NLRI length 255 invalid (> 32)]
Second parameter of bgp_add_route is called "prefixlen" so 0xffffffff will produce the « 255 » length.
( probably from the time when masks where used )
The value should be "32"
Patch to come :)domidomihttps://code.ffdn.org/l2tpns/l2tpns/-/issues/8Clean auto-generated markdown man sources2021-02-05T13:38:13ZJulien RabierClean auto-generated markdown man sourcesCurrent manpages sources in markdown have been generated from the old manuals using Pandoc. The automatic formatting is far from perfect.Current manpages sources in markdown have been generated from the old manuals using Pandoc. The automatic formatting is far from perfect.2.3.3Julien RabierJulien Rabierhttps://code.ffdn.org/l2tpns/l2tpns/-/issues/4Convert docbook documentation to something more user-friendly2021-01-31T17:01:42ZsebianConvert docbook documentation to something more user-friendlyThe actual documentation and manpages are managed by docbook and the `Docs/manual/manual.xml` file.
XML is not very human friendly…
We should maybe use somthing like markdown or asciidoc and use pandoc to convert the documentation (html...The actual documentation and manpages are managed by docbook and the `Docs/manual/manual.xml` file.
XML is not very human friendly…
We should maybe use somthing like markdown or asciidoc and use pandoc to convert the documentation (html for the website, and manpages).2.3.2Julien RabierJulien Rabierhttps://code.ffdn.org/l2tpns/l2tpns/-/issues/3Restarting l2tpns only shuts it down2020-11-07T17:59:44ZzorunRestarting l2tpns only shuts it downWhen trying to restart l2tpns, it fails immediately:
```
# systemctl restart l2tpns
Job for l2tpns.service failed because the control process exited with error code
```
What actually happens is that l2tpns is trying to shut down, but i...When trying to restart l2tpns, it fails immediately:
```
# systemctl restart l2tpns
Job for l2tpns.service failed because the control process exited with error code
```
What actually happens is that l2tpns is trying to shut down, but it can take several seconds, with many messages like this in the logs:
```
...
00/00 Tunnel 5 still has un-acked control messages.
...
00/00 Tunnel 2 still has un-acked control messages.
...
Kill tunnel 2: Timeout on control message
Shutdown complete
```
At this point, l2tpns has exited, and it is not restarted.
What I think happens is that systemd calls "/etc/init.d/l2tpns stop", which exits immediately with a success code, and then calls "/etc/init.d/l2tpns start", which fails because l2tpns is still running:
```
# systemctl status l2tpns.service
● l2tpns.service
Loaded: loaded (/etc/init.d/l2tpns; generated; vendor preset: enabled)
Active: failed (Result: exit-code) since Sun 2020-11-01 19:08:31 CET; 8min ago
Docs: man:systemd-sysv-generator(8)
Process: 15137 ExecStop=/etc/init.d/l2tpns stop (code=exited, status=0/SUCCESS)
Process: 15143 ExecStart=/etc/init.d/l2tpns start (code=exited, status=1/FAILURE)
Tasks: 0 (limit: 4915)
CGroup: /system.slice/l2tpns.service
systemd[1]: Starting l2tpns.service...
l2tpns[15143]: Starting l2tpns:
systemd[1]: l2tpns.service: Control process exited, code=exited status=1
systemd[1]: Failed to start l2tpns.service.
systemd[1]: l2tpns.service: Unit entered failed state.
systemd[1]: l2tpns.service: Failed with result 'exit-code'.
```
A solution would be to switch a proper systemd service file.2.3.0https://code.ffdn.org/l2tpns/l2tpns/-/issues/1Update accounting default directory2020-05-02T16:25:54ZJulien RabierUpdate accounting default directoryOn our setup (Ilico), we encountered an issue where /run was full because of the accounting default setting (/var/run/l2tpns/acct, see [](https://code.ffdn.org/l2tpns/l2tpns/-/blob/master/etc/startup-config.default#L54)).
I think we sho...On our setup (Ilico), we encountered an issue where /run was full because of the accounting default setting (/var/run/l2tpns/acct, see [](https://code.ffdn.org/l2tpns/l2tpns/-/blob/master/etc/startup-config.default#L54)).
I think we should log this somewhere else as /run is volatile (tmpfs) and manage files through logrotate or a cron script.
I'm opening this issue as a reminder to fix this.2.3.0Julien RabierJulien Rabier