[v0.20.3]

#1

hi, All
v0.20.3 发布了！从 v0.20.2 升级的话直接替换掉 bin 文件就可以，详细细节查看 https://docs.citahub.com/zh-CN/cita/protocol-upgrade/overview

Release note:

Fix the issue about high CPU usage caused by too many sst files.
Check the details at #206

修复了由于 sst 文件过多导致的 CPU 占用过高的问题。
更多细节请查看 #206

2赞
[v0.21.1]
#2

升级后跪了。。。
步骤：
1、下载0.20.3的 tar 包，解 tar，覆盖原cita 目录，ls 查看确认文件确实已被更新；
2、依次执行 ./env.sh ./bin/cita stop casewhen-testchain/x 停止每个节点；
3、依次执行 ./daemon.sh ./bin/cita start casewhen-testchain/x 启动每个节点；
4、查看日志发现启动跪了；

报错日志如下，似乎是说 rabbitmq 没开启？

============================

2019-03-12T04:11:12.009269903+00:00 - INFO - CITA:executor
2019-03-12T04:11:12.009319464+00:00 - INFO - Version: v0.20.2-1-g0eb18fe
2019-03-12T04:11:12.030298915+00:00 - ERROR -
============================
stack backtrace:
   0:     0x55c1867f40ec - backtrace::backtrace::trace::hfa99d66440780557
   1:     0x55c1867f3362 - <backtrace::capture::Backtrace as core::default::Default>::default::h4d3bf2e7afd53899
   2:     0x55c1867f33d8 - backtrace::capture::Backtrace::new::h92a28e62d17e981a
   3:     0x55c1867f2688 - panic_hook::panic_hook::h733eea3f492eb849
   4:     0x55c1867f23c8 - core::ops::function::Fn::call::h9b40aff6656fca0a
   5:     0x55c186b74fc3 - std::panicking::rust_panic_with_hook::hd470ff3b3e7a1b18
                        at libstd/panicking.rs:480
   6:     0x55c186b74b29 - std::panicking::continue_panic_fmt::hf0cf39ae1e114602
                        at libstd/panicking.rs:390
   7:     0x55c186b74a8d - std::panicking::begin_panic_fmt::hfc5862442bd06bfc
                        at libstd/panicking.rs:345
   8:     0x55c18636b82e - pubsub_rabbitmq::start_rabbitmq::hb1a3122dbfb26a55
   9:     0x55c1863358c2 - pubsub::start_pubsub::h8f4f7a7812174300
  10:     0x55c186363f9b - cita_executor::main::h954ea2b9b97d9986
  11:     0x55c186351952 - std::rt::lang_start::{{closure}}::h82636323630ad674
  12:     0x55c186b749c2 - std::rt::lang_start_internal::{{closure}}::h444bf05eb286eca1
                        at libstd/rt.rs:59
                         - std::panicking::try::do_call::hb527d54965a8815a
                        at libstd/panicking.rs:310
  13:     0x55c186b84bd9 - __rust_maybe_catch_panic
                        at libpanic_unwind/lib.rs:102
  14:     0x55c186b64105 - std::panicking::try::h83b5294aab121d1a
                        at libstd/panicking.rs:289
                         - std::panic::catch_unwind::h15c1ea783c7910d5
                        at libstd/panic.rs:392
                         - std::rt::lang_start_internal::hcc3e9196a926118d
                        at libstd/rt.rs:58
  15:     0x55c186365493 - main
  16:     0x7f0a1b2ddb96 - __libc_start_main
  17:     0x55c186319359 - _start
  18:                0x0 - <unknown>

position:
Thread main panicked at failed to open url amqp://guest:guest@localhost/casewhen-testchain/2 : IoError(ConnectionRefused), /opt/.cargo/git/checkouts/cita-common-1aad419f3e80ba17/8d37eb7/pubsub_rabbitmq/src/lib.rs:59

This is a bug. Please report it at:

    https://github.com/cryptape/cita/issues/new

============================
#3

覆盖整个 cita 目录吗还是 bin 目录？
看日志怎么还是 v0.20.2？

image 删掉重新 setup/start 试下

#4

覆盖了整个 cita 目录；
重新 setup 后，start 成功了

#5

这样数据就丢失了呀，只需要把 bin 目录覆盖i就行了。

这样相当于重新起了个新的链，不算升级了

#6

并没有，下载的tar 包解压后不包含我建链的目录，所以实际上就是除了链目录以外的其他所有文件都被覆盖率，数据没有丢失
image

#7

Got it.

不过按照道理来说，rabbitmq 不用重新 setup 的