? Editing: Post:21.body Save Delete Cancel
Content changed Sign & Publish new content

ZeroBlog

Demo for decentralized, self publishing blogging platform.

Follow in NewsfeedFollowing

Latest comments:

Add new post

Title

21 hours ago · 2 min read ·
3 comments
Body
Read more

Not found

Changelog: August 28, 2018

on Aug 28, 2018

Shared tracker list between users:

New tracker IPs automatically discovered from other users to reduce centralization and avoid the blocking of the trackers.

To start new tracker: Enable the Bootstrapper plugin with opened port, then it will automatically share your client's ip with other users as a possible tracker. (It's only recommended to do this if you have static ip and planning to keep your client running 24/7)

Only the working trackers (successful result in past hour) are shared with other peers and the discovery request done in every 5 minutes until your client found 5 working shared trackers. This feature is also got limited to zero:// trackers only as it more suited for ZeroNet, because it allows multiple sites announce in same request and the storage of .onion addresses.

Other changes (Rev3571):

  • Copy site address with peers link on the sidebar.

  • Zip file listing and streaming support for Bigfiles.

  • Tracker statistics on /Stats page

  • Allow tracker connections to loopback to your client and announce yourself.

  • Reduce the request frequency to unreliable trackers.

  • Fix double announce on new site visit and startup.

  • Fix trayicon compatibility with latest gevent

  • Fix request number counting for zero:// trackers

14 Comments:

user_name1 day ago
Reply
Body
syntiaon May 03, 2019
Reply

Толковые хлопцы!

gsdthon Mar 08, 2019
Reply

super

burningsserveron Sep 21, 2018
Reply

cool post

solenson Sep 05, 2018
Reply

ty

progion Sep 03, 2018
Reply

The system is better from day to day. :)

Some positive feedback:

1.The tracker system: fast connection, the users quickly find each other.

  1. Some zites finaly downloaded later the next day what i want too see.I do not have to try again and again.Once it appears in the zite list. This is a huge step positive forward.
  2. Big files transmission are much better.

What I find functionally the most important in this system as website administrator. If a user want a file or zite download. And the user who have the file or zite are not online. But later will be online. At this point the downloader automatically get the file or zite. But especially with zites. I see great progress in this area. The most important if someone wants to download something, all the same its a zite or file. If someone starts seed the content in this case the user get this content.And do not have to constantly attempt to unsuccessfully try to get the files and zites.The cost of maintaining the website,bandwidth,zite uptime (without 24/7 run my PC),the users successfully get or not what they like,the number of users and seeders,In short, the most important thing.

nomuon Sep 02, 2018
Reply

Nice

geeklesson Aug 29, 2018
Reply

nofish: I have added a modification that is limits the concurrent siteCheck calls to 5 and also reduced the opened connection number for startup.

Thank you!

nofishon Aug 29, 2018
Reply

geekless: @nofish:
I have 500+ zites hosted on my computer and I use the Tor Always mode. When I start Zeronet, it opens a really HUGE number of connections, so the tor daemon gets unresponsive for a long time, and finally all the connections are failed with timeout. It is kind of self-DoS.

Thanks for the suggestion, I have added a modification that is limits the concurrent siteCheck calls to 5: https://github.com/HelloZeroNet/ZeroNet/commit/3a9ded9315f6e3ec4543ce6c464c38172dcbecba and also reduced the opened connection number for startup.

usersdoctoron Aug 29, 2018
Reply

+1

maxiaoon Aug 29, 2018
Reply

+1

jadon Aug 29, 2018
Reply

It's great! Thank you!+1

geeklesson Aug 28, 2018
Reply

@nofish:

Please consider applying the following patch:

diff --git a/src/File/FileServer.py b/src/File/FileServer.py
index 0e4a8a7..cb78a99 100644
--- a/src/File/FileServer.py
+++ b/src/File/FileServer.py
@@ -277,11 +277,20 @@ class FileServer(ConnectionServer):
                 self.tor_manager.startOnions()

         if not sites_checking:
+            check_threads = []
             for site in sorted(self.sites.values(), key=lambda site: site.settings.get("modified", 0), reverse=True):  # Check sites integrity
+                self.log.debug("Checking site: %s" % site.address_short)
                 check_thread = gevent.spawn(self.checkSite, site, check_files)  # Check in new thread
-                time.sleep(2)
-                if site.settings.get("modified", 0) < time.time() - 60 * 60 * 24:  # Not so active site, wait some sec to finish
-                    check_thread.join(timeout=5)
+                check_threads.append(check_thread)
+                i = 0
+                while 1:
+                    check_threads = [x for x in check_threads if not x.ready()]
+                    self.log.debug("check_threads: %s" % len(check_threads))
+                    if len(check_threads) < 10 and i > len(check_threads):
+                        break
+                    check_threads[0].join(timeout=1)
+                    i += 1
+

     def cleanupSites(self):
         import gc

I have 500+ zites hosted on my computer and I use the Tor Always mode. When I start Zeronet, it opens a really HUGE number of connections, so the tor daemon gets unresponsive for a long time, and finally all the connections are failed with timeout. It is kind of self-DoS.

I modified the checkSites() method to implement a simple self-balancing algorithm. It uses a variable time interval before queuing the next zite in the list, depending on the current queue lengh, and the maximum lengh is limited by 10.

kaffieon Aug 28, 2018
Reply

Sounds amazing. But when can we get IPv6 support for all this?

geeklesson Aug 28, 2018
Reply

It's great! Thank you!

This page is a snapshot of ZeroNet. Start your own ZeroNet for complete experience. Learn More