Author Topic: 【翻译贴】第六轮测试总结和第七轮测试展望  (Read 12685 times)

0 Members and 1 Guest are viewing this topic.

Offline PTS中国

  • Sr. Member
  • ****
  • Posts: 416
    • View Profile
  • BitShares: ptschina
:'( :'( :'( alt 哥
熊熊这是喜极而泣了吗? :)

二徒弟为大徒弟喝彩吧,他呀,终于要逃离桑拿天的高老庄了
--------

PTS中国

Offline alt

  • Hero Member
  • *****
  • Posts: 2821
    • View Profile
  • BitShares: baozi

Offline Musewhale

  • Hero Member
  • *****
  • Posts: 2881
  • 丑,实在是太丑了 !
    • View Profile
MUSE witness:mygoodfriend     vote for me

Offline crazybit


Offline HackFisher

  • Hero Member
  • *****
  • Posts: 883
    • View Profile
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline ebit

  • Committee member
  • Hero Member
  • *
  • Posts: 1905
    • View Profile
  • BitShares: ebit
我好像找到了不产生区块的bug  :)

看来这一轮进步很大啊,同学们,东方已经鱼肚白了,就等着欢呼吧 :)

走出黑暗,迎接光明
telegram:ebit521
https://weibo.com/ebiter

Offline mitao

  • Full Member
  • ***
  • Posts: 113
    • View Profile
非常确信用不了几天了


Sent from my iPhone using Tapatalk

Offline yinchanggong

  • Sr. Member
  • ****
  • Posts: 464
    • View Profile
    • 微博 引长弓Fate
BTSX delegate: google.helloworld    microsoft.helloworld
BTSX Account:yinchg   Manager of BTSXCHINA Charity Fund
引长弓Fate 新浪微博

Offline PTS中国

  • Sr. Member
  • ****
  • Posts: 416
    • View Profile
  • BitShares: ptschina
我好像找到了不产生区块的bug  :)

看来这一轮进步很大啊,同学们,东方已经鱼肚白了,就等着欢呼吧 :)
--------

PTS中国

Offline alt

  • Hero Member
  • *****
  • Posts: 2821
    • View Profile
  • BitShares: baozi
我好像找到了不产生区块的bug  :)

Offline baijie

  • Jr. Member
  • **
  • Posts: 39
    • View Profile

Offline alt

  • Hero Member
  • *****
  • Posts: 2821
    • View Profile
  • BitShares: baozi
看进度,比较欣慰,大家加油!

Quote
代理不生产区块的问题
,就是
Quote
2148,{
      "missed": true,
      "id": "0000000000000000000000000000000000000000"
吧?

非代表应该觉察不到吧?这个代表不产生区块,其他代表就代替他去产生区块了吧?
其它代表是不会代替的,就直接错过了,比如2393,2394之间漏过了两个区块,就间隔了6分钟。

Offline ebit

  • Committee member
  • Hero Member
  • *
  • Posts: 1905
    • View Profile
  • BitShares: ebit
看进度,比较欣慰,大家加油!

Quote
代理不生产区块的问题
,就是
Quote
2148,{
      "missed": true,
      "id": "0000000000000000000000000000000000000000"
吧?

非代表应该觉察不到吧?这个代表不产生区块,其他代表就代替他去产生区块了吧?
« Last Edit: July 01, 2014, 01:50:28 pm by ebit »
telegram:ebit521
https://weibo.com/ebiter

Offline bennyliaa

  • Full Member
  • ***
  • Posts: 64
    • View Profile
My Personal BTS Account: stephen

Offline alt

  • Hero Member
  • *****
  • Posts: 2821
    • View Profile
  • BitShares: baozi
Quote
45 forks is not normal operation and is a result of bugs in the networking code网络代码.  I will have Eric look into the cause when he comes in on Monday.

Quote
The issue with forking is not because of slow systems or network problems.  It is almost entirely due to implementation bugs.  I am fairly certain we have narrowed it down to a blocking socket write that prevents other sockets from being sent data until a TCP timeout.


这些问题用“优化数据库”和调节有效块的方式能一劳永逸的解决么?
目前系统比较稳定了,但仍然不让人放心啊。
这轮测试的一个收获是 fork switch 比较稳定了,也就是当产生分叉后,都能正确的切换到主链上,这很重要。
主要的问题是生产区块效率太低,包括:区块广播延迟大,因为延迟又导致分叉,还有很多代理没正常产生区块。
今天的测试结果,网络延迟似乎已经解决了,可用以下命令查看,我看到的延迟最大是3秒。
Code: [Select]
delegate (unlocked) >>> blockchain_list_blocks 1960 800
分叉似乎也没有再产生了,我这看到最后的分叉是1866, 从1960块之后没有分叉
Code: [Select]
delegate (unlocked) >>> blockchain_list_forks

代理不生产区块的问题还没解决,bm好像正在处理。

Offline gyhy

  • Hero Member
  • *****
  • Posts: 852
    • View Profile
Quote
45 forks is not normal operation and is a result of bugs in the networking code网络代码.  I will have Eric look into the cause when he comes in on Monday.

Quote
The issue with forking is not because of slow systems or network problems.  It is almost entirely due to implementation bugs.  I am fairly certain we have narrowed it down to a blocking socket write that prevents other sockets from being sent data until a TCP timeout.


这些问题用“优化数据库”和调节有效块的方式能一劳永逸的解决么?
目前系统比较稳定了,但仍然不让人放心啊。
慢慢优化吧

Offline ebit

  • Committee member
  • Hero Member
  • *
  • Posts: 1905
    • View Profile
  • BitShares: ebit
Quote
45 forks is not normal operation and is a result of bugs in the networking code网络代码.  I will have Eric look into the cause when he comes in on Monday.

Quote
The issue with forking is not because of slow systems or network problems.  It is almost entirely due to implementation bugs.  I am fairly certain we have narrowed it down to a blocking socket write that prevents other sockets from being sent data until a TCP timeout.


这些问题用“优化数据库”和调节有效块的方式能一劳永逸的解决么?
目前系统比较稳定了,但仍然不让人放心啊。
telegram:ebit521
https://weibo.com/ebiter


Offline suwoder

  • Sr. Member
  • ****
  • Posts: 206
    • View Profile
谢天谢地,最好测试不要超过10轮,不然论坛又要爆发了 ;)

Offline rnixianren

  • Full Member
  • ***
  • Posts: 114
    • View Profile
不管以前对3I有多大怨气,他们现在确实是在认真做事情了!要是以前一直都这个样子就好了,说不定大家早就小赚一点了

Offline dcchong

  • Sr. Member
  • ****
  • Posts: 203
    • View Profile
wallet_approve_delegate dc-delegate true
wallet_approve_delegate bitsharesx-delegate true

Offline Overthetop

不是被盗号了, :)而是 你们一直误解我。就像任志强一样。

哎呀,我的妈呀
把我笑死了今天 :-X :-X :-X

看来有人弃暗投明啦  :D
个人微博账号: Overthetop_万里晴空
“块链创新与创业”交流群: 330378613

Offline Overthetop

个人微博账号: Overthetop_万里晴空
“块链创新与创业”交流群: 330378613

Offline suzhu01

  • Sr. Member
  • ****
  • Posts: 203
    • View Profile
不是被盗号了, :)而是 你们一直误解我。就像任志强一样。

哎呀,我的妈呀
把我笑死了今天 :-X :-X :-X
肯定被盗号了
PTS:PfshcXbRBQ1ouepXYN4QG1kjTLgvtPL68W
BTSX Account:suzhu01

Offline Musewhale

  • Hero Member
  • *****
  • Posts: 2881
  • 丑,实在是太丑了 !
    • View Profile
不是被盗号了, :)而是 你们一直误解我。就像任志强一样。

哎呀,我的妈呀
把我笑死了今天 :-X :-X :-X
MUSE witness:mygoodfriend     vote for me

Offline ripplexiaoshan

  • Moderator
  • Hero Member
  • *****
  • Posts: 2300
    • View Profile
  • BitShares: jademont
不是被盗号了, :)而是 你们一直误解我。就像任志强一样。
+5% +5% +5%
BTS committee member:jademont

Offline hero

  • Jr. Member
  • **
  • Posts: 45
    • View Profile
楼上的被盗号了??
真有喜感

来自我的 Nexus 5 上的 Tapatalk


Offline gyhy

  • Hero Member
  • *****
  • Posts: 852
    • View Profile
这确实是一个复杂的分布式系统。要处理好多情况

Offline zhao150

  • Hero Member
  • *****
  • Posts: 606
  • 老子早就不想当代表了
    • View Profile
不是被盗号了, :)而是 你们一直误解我。就像任志强一样。
老子早就不想当代表了


Offline PTS中国

  • Sr. Member
  • ****
  • Posts: 416
    • View Profile
  • BitShares: ptschina
--------

PTS中国

Offline dexinwong

  • Sr. Member
  • ****
  • Posts: 232
    • View Profile

Offline zhao150

  • Hero Member
  • *****
  • Posts: 606
  • 老子早就不想当代表了
    • View Profile
看上去不错,测试的好
其实 我是非常欢迎3i 搞一轮一轮的测试,非常严谨,像是搞金融开发的料。
其他有些技术含量的山寨币,没有经过像3i一样严密的测试,就发布了。
那样真的不太负责
老子早就不想当代表了

Offline Snail

  • Hero Member
  • *****
  • Posts: 750
    • View Profile
  • BitShares: snail

Offline ebit

  • Committee member
  • Hero Member
  • *
  • Posts: 1905
    • View Profile
  • BitShares: ebit
telegram:ebit521
https://weibo.com/ebiter


Offline metalallen

  • Sr. Member
  • ****
  • Posts: 262
    • View Profile
BM原帖:https://bitsharestalk.org/index.php?topic=5303.msg71585#msg71585

译文:

截止世界标准时间4:46,最后一个分叉是在第1866区块上。Eric做了程序检查并在修复中,他应该可以消除导致延迟和分叉的一个主要原因。

我们已经确定了以下问题有待解决:
1、在我们整个周末的长时间压力测试影响下,整个网络同步状况异常缓慢。
     - 我们会将数据库优化,应该可以解决这个问题。

2、扫描区块链中的属于你自己的交易信息,会占用电脑大量的CPU资源。
      - 我们有一个针对性的整体架构设计调整已经到位,将会给大家提供轻量级的运行支持。虽然调整了架构,但这并不会推迟发布,因为我们不希望在推出轻客户端(也许很久之后)之前的这段时间让每秒两次交易(频率太低)成为常态。

3、我们会在48小时之后让现在所有交易数据过期。
      - 这会消除数据库中的一个重大问题,并会减少验证区块的完整节点所需的存储量;
      - 这会最终让同步仅“当前所在时区”+48小时的区块链内交易成为可能;
      - 这会是个小变化,但会导致硬分叉。(译者注:不必担心,新一轮测试可能沿用分叉后的新链)

4、我们将会允许普通账户升级成为受托人,和降低交易手续费。
      - 这会有硬分叉(不必担心),会实现确认48小时的交易。

新一轮测试将于明天晚上开始,届时我们会修复所有已知的bug,并且希望在第七轮测试时彻底让分叉问题终结。目前,请受托人们不要停止出块(使用最新客户端),我们将在下一轮测试前观察分叉是否在减少。
浮壹白的微博:http://weibo.com/u/2279693077
BTSX Account:metalallen