I have filtered most problematic trackers out of my torrents using the options at Settings/Transfers/Trackers

In making some adjustments today, I was reminded of a minor matter regarding the handling of dead trackers.

I saw some reported as "Host Not Found", with a try again set for 2 minutes.

My experience is that, if the system is function and other trackers are working normally, one reporting "Host Not Found" is dead, likely an abandoned domain name, never to be returned to tracker service.

If that is a fair assessment, then the question is, why try probing it every 2 minutes. Should there be an other behavior configured? How about once every 24 hours?  And how about logging the last time connected and if it goes past 30/60/90? days as "Host Not Found", then auto-add it to the skip list with a note?

Bad idea. Trackers go down then come back. A couple of them went down for over a week not long ago. I would not want to have them on the list. Some jobs are meant for the user, this is one. The other trackers that need to be added to that list are the ones giving out IPs. These trackers should all go in the remove column.

There are many more that could go on the list, but these are proven to divulge the IPs of users.
Hope this is OK as they are not torrent sites but trackers.


(EDITED BY MOD--list of trackers removed)

Just to bud in, I agree with the OP. Reply #1 states OPs idea as bad because some times they go down. Then gives an example with a week of downtime. The OP was talk 30/60/90 days... I think the OP has a good and well  pointed out argument.

Just my view :)

I can see several things like "connection refused? and "Timed out" as perhaps being temporary. The server is registered on DNS, perhaps running, but not offering the expected service.


But I think "Host Not Found" is a different category of problem in that it is no longer registered with the DNS network. Certainly there is no need to keep trying every 2 minutes.

Were the ones that you noticed going down and coming back within a couple weeks "Host not found" or some other issue preventing connection?

My suggestion was merely to add these to the skip list. And if my idea of a manual announce override was also implemented, then when the user sees a tracker being skipped, they could force a manual announce to test the current status and if it works, then that tracker could be removed from the skip list. As is now, manually announcing to a skipped tracker does nothing except report it will still be skipped. I think a manual announce should override the skip list and let the test go through.

Host not found is exactly that. Sometimes it happens for an hour. Both examples were suppose to be down for over a month or more to reset server setups. I just went though my trackers and am showing 4 that are not found. 3 I know will be found shortly.

For an idea to be good we have to ask what harm is there? How much do we need to add to the proggy to make it work.

An easy way to stop this and in general a good idea: go to settings > transfers > trackers and change Multi-tracker announce mode to First working group. This setting makes Tixati connect to the first working tracker and not waste bandwidth on other trackers. This + DHT will work great.