feature(*): support exec subcommand

[aspirer]

issue #153
usage:
curveadm exec 0e5d809f16b4 ps
curveadm exec 0e5d809f16b4 curve_ops_tool status
curveadm exec 0e5d809f16b4 -- ps -ef

Signed-off-by: Wangpan [email protected]

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec -h
Usage:  curveadm exec ID [OPTIONS]

Exec a cmd in service container

Options:
  -h, --help   Print usage

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 
Connection to 127.0.0.1 closed.
netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 -h
Usage:  curveadm exec ID [OPTIONS]

Exec a cmd in service container

Options:
  -h, --help   Print usage

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 ps
    PID TTY          TIME CMD
   1652 pts/3    00:00:00 ps
Connection to 127.0.0.1 closed.

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 "ps -ef"
UID          PID    PPID  C STIME TTY          TIME CMD
root           1       0  0 Nov17 ?        00:00:01 /sbin/docker-init -- /entrypoint.sh --role mds --args=

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 -- ps -ef
UID          PID    PPID  C STIME TTY          TIME CMD
root           1       0  0 Nov17 ?        00:00:01 /sbin/docker-init -- /entrypoint.sh --role mds --args=

netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 curve_ops_tool status | head -n3
Cluster status:
Send RPC to chunkserver fail, error content: [E112]Not connected to 127.0.0.1:8200 yet, server_id=6 [R1][E112]Not connected to 127.0.0.1:8200 yet, server_id=6 [R2][E112]Not connected to 127.0.0.1:8200 yet, server_id=6 [R3][E112]Not connected to 127.0.0.1:8200 yet, server_id=6
Send RPC to chunkserver fail, error content: [E112]Not connected to 127.0.0.1:8202 yet, server_id=8 [R1][E112]Not connected to 127.0.0.1:8202 yet, server_id=8 [R2][E112]Not connected to 127.0.0.1:8202 yet, server_id=8 [R3][E112]Not connected to 127.0.0.1:8202 yet, server_id=8
Connection to 127.0.0.1 closed.
netease@hih-l-9493:~/curveadm$ ./bin/curveadm exec 0e5d809f16b4 -- curve_ops_tool status | head -n3
Cluster status:

[Cyber-SiKu]

Is curveadm exec [--id xxxx] cmd better? The id can be omitted and it will be executed on the main mds by default.

[Cyber-SiKu]

name

[Cyber-SiKu]

ditto

[Cyber-SiKu]

2022

[Wine93]

How about omit the explicit cmd argument, like:

curveadm exec 0e5d809f16b4 ps -ef

it's more like docker's exec command:

docker exec 4ac34856f43f ps -ef

[aspirer]

How about omit the explicit cmd argument, like:

curveadm exec 0e5d809f16b4 ps -ef

it's more like docker's exec command:

docker exec 4ac34856f43f ps -ef

usage:
curveadm exec 0e5d809f16b4 ps
curveadm exec 0e5d809f16b4 curve_ops_tool status
curveadm exec 0e5d809f16b4 -- ps -ef