Unverified 提交 92c88b0a 作者: Steven Allen 提交者: GitHub

Merge pull request #6268 from ipfs/doc/environment

docs: document environment variables
# go-ipfs environment variables
## `LIBP2P_TCP_REUSEPORT` (`IPFS_REUSEPORT`)
go-ipfs tries to reuse the same source port for all connections to improve NAT
traversal. If this is an issue, you can disable it by setting
`LIBP2P_TCP_REUSEPORT` to false.
This variable was previously `IPFS_REUSEPORT`.
Default: true
## `IPFS_PATH`
Sets the location of the IPFS repo (where the config, blocks, etc.
are stored).
Default: ~/.ipfs
## `IPFS_LOGGING`
Sets the log level for go-ipfs. It can be set to one of:
* `CRITICAL`
* `ERROR`
* `WARNING`
* `NOTICE`
* `INFO`
* `DEBUG`
Logging can also be configured (on a subsystem by subsystem basis) at runtime
with the `ipfs log` command.
Default: `ERROR`
## `IPFS_LOGGING_FMT`
Sets the log message format. Can be one of:
* `color`
* `nocolor`
Default: `color`
## `GOLOG_FILE`
Sets the file to which go-ipfs logs. By default, go-ipfs logs to standard error.
## `GOLOG_TRACING_FILE`
Sets the file to which go-ipfs sends tracing events. By default, tracing is
disabled.
This log can be read at runtime (without writing it to a file) using the `ipfs
log tail` command.
Warning: Enabling tracing will likely affect performance.
## `IPFS_FUSE_DEBUG`
Enables fuse debug logging.
Default: false
## `YAMUX_DEBUG`
Enables debug logging for the yamux stream muxer.
Default: false
## `IPFS_FD_MAX`
Sets the file descriptor limit for go-ipfs. If go-ipfs fails to set the file
descriptor limit, it will log an error.
Defaults: 2048
## `IPFS_DIST_PATH`
URL from which go-ipfs fetches repo migrations (when the daemon is launched with
the `--migrate` flag).
Default: https://ipfs.io/ipfs/$something (depends on the IPFS version)
## `LIBP2P_MUX_PREFS`
Tells go-ipfs which multiplexers to use in which order.
Default: "/yamux/1.0.0 /mplex/6.7.0"
...@@ -104,13 +104,12 @@ bandwidth than the others, but is lacking on congestion control and backpressure ...@@ -104,13 +104,12 @@ bandwidth than the others, but is lacking on congestion control and backpressure
logic. It is available to try out and experiment with. logic. It is available to try out and experiment with.
### State ### State
Experimental Stable
### In Version ### In Version
0.4.5 0.4.5
### How to enable ### How to enable
run your daemon with `--enable-mplex-experiment`
To make it the default stream muxer, set the environment variable To make it the default stream muxer, set the environment variable
`LIBP2P_MUX_PREFS` as follows: `LIBP2P_MUX_PREFS` as follows:
...@@ -118,37 +117,6 @@ To make it the default stream muxer, set the environment variable ...@@ -118,37 +117,6 @@ To make it the default stream muxer, set the environment variable
export LIBP2P_MUX_PREFS="/mplex/6.7.0 /yamux/1.0.0 /spdy/3.1.0" export LIBP2P_MUX_PREFS="/mplex/6.7.0 /yamux/1.0.0 /spdy/3.1.0"
``` ```
To check which stream muxer is being used between any two given peers, check the
json output of the `ipfs swarm peers` command, you'll see something like this:
```
$ ipfs swarm peers -v --enc=json | jq .
{
"Peers": [
{
"Addr": "/ip4/104.131.131.82/tcp/4001",
"Peer": "QmaCpDMGvV2BGHeYERUEnRQAwe3N8SzbUtfsmvsqQLuvuJ",
"Latency": "46.032624ms",
"Muxer": "*peerstream_multiplex.conn",
"Streams": [
{
"Protocol": "/ipfs/bitswap/1.1.0"
},
{
"Protocol": "/ipfs/kad/1.0.0"
},
{
"Protocol": "/ipfs/kad/1.0.0"
}
]
},
{
...
```
### Road to being a real feature
- [ ] Significant real world testing and performance metrics across a wide
variety of workloads showing that it works well.
--- ---
## Raw Leaves for unixfs files ## Raw Leaves for unixfs files
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论