Hi,

On 9 Mar 2019, at 13:08, Roger Dingledine <arma@torproject.org> wrote:

...

I also have some questions.

Atlas does not show any IPv6 address, is this normal?

I see that you have an ipv6 exit policy set, but I don't see any ipv6 address in your relay descriptor.

(You can see your relay descriptor with "wget 5.199.130.188/tor/server/authority" )

If you were advertising an ipv6 address, you would have an "or-address" line in your descriptor. Compare to Fission1's descriptor: "wget 158.69.30.132/tor/server/authority"

If you want to set an IPv6 address, use: ORPort [IPv6]:Port For example: ORPort [2001:db8::1]:9001 Tor doesn't guess IPv6 addresses yet. T

Hi,

On 12 Mar 2019, at 05:34, ylms <tor@yl.ms> wrote:

...

On 3/9/19 5:07 AM, teor wrote: ORPort [IPv6]:Port

For example:

ORPort [2001:db8::1]:9001

Tor doesn't guess IPv6 addresses yet.

That was very helpful to find my stupid mistake.

We would like to make tor more helpful, and make it guess IPv6. But we're not there yet. Until then, people have to learn the details of tor's config :-( T

Hello Roger, thanks for you reply. On 3/9/19 4:08 AM, Roger Dingledine wrote:

On Fri, Mar 08, 2019 at 03:52:48PM +0100, ylms wrote:

...

Hello, I just setup a new exit, I know it takes time to "train", but I would be interested in tips on how to tweak it. I just replaces a old exit that used the same IP, but unfortunately we lost that config/key.

So any tips are welcome in general.

It looks good so far! Thanks for running it. It is good to see some good news from your part of Germany, not just bad proposed policies. :)

Yeah, as said it is not a new relay, just lost the old one. But I would like to add a new one, just finding a useful Germany hoster is not easy to find.

...

I also have some questions.

Atlas does not show any IPv6 address, is this normal?

I see that you have an ipv6 exit policy set, but I don't see any ipv6 address in your relay descriptor.

(You can see your relay descriptor with "wget 5.199.130.188/tor/server/authority" )

If you were advertising an ipv6 address, you would have an "or-address" line in your descriptor. Compare to Fission1's descriptor: "wget 158.69.30.132/tor/server/authority"

I fixed it. I did not think about the OR-Port but did set: OutboundBindAddress 5.199.130.188 # Use this IP for all outgoing connections OutboundBindAddress [2001:4ba0:fff9:160:dead:beef:ca1f:1337] # Use this IP for all outgoing connections OutboundBindAddressOR 5.199.130.188 # OR Port IP OutboundBindAddressOR [2001:4ba0:fff9:160:dead:beef:ca1f:1337] # OR Port IP OutboundBindAddressExit 5.199.130.188 # Exit IP OutboundBindAddressExit [2001:4ba0:fff9:160:dead:beef:ca1f:1337] # Exit IP And for some reason I thought "Address" was the setting to do so, so I had: Address 5.199.130.188 and Address [2001:4ba0:fff9:160:dead:beef:ca1f:1337] But I removed that because it was bogus, so now it is: Address tor.piratenpartei-nrw.de now, which is what I meant to set.

...

WHat loglevel can I set to see if there is any problems, to find possible improvements? I know the logging should not be changed, but I want tot make sure tor works fine, so I would like to see some more information for now.

We recommend notice-level logs. That's what most of the Tor packages do by default: https://www.torproject.org/docs/faq#LogLevel

So "err" or "warn" would be the choice for troubleshooting, I assume this can be changed and a "systemctl reload tor.service" would be enough? Thanks for the help again. yl

Thanks! --Roger

_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays

On 2019-04-05 11:31, lists@for-privacy.net wrote:

Am 05.04.2019 10:58, schrieb ylms:

...

can someone point me at some information about this warning?

"[WARN] Error binding network socket: Address already in use [991 duplicates hidden]"

Log message is clear: You have assigned a port number twice. Either two Tor-instances run on the same (TCP) port numbers or you have given Tor a port number, which already has another system process.

'netstat -lptu' or successor 'ss -lptu' gives you an overview.

One thing that could happen (but may not apply to you) is your SOCKSPort. Set it to 0 to not listen, or to a random number if you have to listen for SOCKS connections. -Neel === https://www.neelc.org/

On 4/5/19 6:16 PM, lists@for-privacy.net wrote:

Hint for ylms: Socks 0 + reject in the torrc config.

SocksPort 0 SocksPolicy reject *

I think the latter is not needed because I did set "SocksPort 0" now, then reloaded the config and got this: 09:06:52 [NOTICE] Closing no-longer-configured Socks listener on /run/tor/socks:0 │ 09:06:52 [NOTICE] Closing no-longer-configured Socks listener on 127.0.0.1:9050 So I guess the second is not needed. But thanks you anyway. yl

On 4/5/19 10:58 AM, ylms wrote:

"[WARN] Error binding network socket: Address already in use [991 duplicates hidden]"

I did use a search engine, but am not sure what too look for further. Also let me know if I should not worry at all.

991 warnings are probably since tor is running, which is about 24 days.

I just realized that the fault did not occur since I asked about the problem here, I did not change anything till today, so maybe there was another condition present at some time before. Also I did do some other checks, see my emails in the this thread, and did not find any problems. yl