startupcmd: some options don't work successfully [development_fixes103]

Dinaamagdy commented 5 years ago

- stop with foreground

Steps

startupcmd_server = j.servers.startupcmd.get() startupcmd_server.executor = "foreground" startupcmd_server.cmd_start = "python3 -m http.server" tmux -d -s session 'kosmos -p "j.servers.startupcmd.get().start()"' tmux -d -s session 'kosmos -p "j.servers.startupcmd.get().stop()"'

Issue

stop return True , but it doesn't stop the running server

- corex executor doesn't work:

startupcmd =j.servers.startupcmd.get("") j.servers.corex.default.start() corex = j.servers.corex.default.client startupcmd.executor = "corex" startupcmd.corex_client_name = corex.name startupcmd.cmd_start="python3 -m http.server"
startupcmd.start()

--TRACEBACK------------------
<stdin> in <module>
    1    
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in start
    500    if not reset and self.is_running() == True:
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in is_running
    336    return self._error_raise("cannot check running don't have pid or corex_id")
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in _error_raise
    90    raise j.exceptions.Base(msg)
-----------------------------
Tue 24 08:55:35 tupcmd/StartupCMD.py -  90 - _error_raise                       : EXCEPTION: 
    error in startupcmd :ee
    cannot check running don't have pid or corex_id
None

- JUMPSCALE interpreter

server=j.servers.startupcmd.get("dinad")
server.interpreter="JUMPSCALE"
server.cmd_start="j.sal.fs.createEmptyFile('/root/dina')" servet.start()

    error in startupcmd :dinad
    could not start
--TRACEBACK------------------
<stdin> in <module>
    1    
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in start
    626    running = self.wait_running(die=True)
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in wait_running
    477    return self._error_raise("could not start")
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in _error_raise
    90    raise j.exceptions.Base(msg)
-----------------------------
![image](https://user-images.githubusercontent.com/24386802/65507707-c74f4100-dece-11e9-9ee5-aca1154343b2.png)

Dinaamagdy commented 5 years ago

corex executor doesn't work:

startupcmd =j.servers.startupcmd.get("") j.servers.corex.default.start() corex = j.servers.corex.default.client startupcmd.executor = "corex" startupcmd.corex_client_name = corex.name startupcmd.cmd_start="python3 -m http.server"
startupcmd.start()


--TRACEBACK------------------
<stdin> in <module>
    1    
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in start
    500    if not reset and self.is_running() == True:
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in is_running
    336    return self._error_raise("cannot check running don't have pid or corex_id")
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in _error_raise
    90    raise j.exceptions.Base(msg)
-----------------------------
Tue 24 08:55:35 tupcmd/StartupCMD.py -  90 - _error_raise                       : EXCEPTION: 
    error in startupcmd :ee
    cannot check running don't have pid or corex_id
None```

ranatrk commented 5 years ago

Issue : stop with foreground The issue occured because the pid was not set for the startupcmd as once the cmd was executed in foreground it was blocking and therefore the rest of the start cmd wasnt executed. When stopping there was no pid set to stop the process by

Fix: Make self.pid a property that is retrieved everytime by getting the pid from the name of the startupcmd and not using the one saved in the innstance itself

ranatrk commented 5 years ago

Issue: JUMPSCALE interpreter Fix: https://github.com/threefoldtech/jumpscaleX_core/commit/2e14a31069aa08810c293cb8cee8bac7f0c858e8

ranatrk commented 5 years ago

WIP: corex executor doesn't work fixes so far: start and stop working for the first time.

The issue with start() was because setting self.pid didnt occur correctly when using corex.
The issue with stop() was because in process property the state is set to running even though it is stopped or stopping because the process itself obtained using j.sal.process.getProcessObject(pid) has a sleeping status

Current issue: When starting then stopping then starting again, the previous corex pid is used and there is no process that is currently running with this pid. Working on resetting corex data when stopping the server

ranatrk commented 5 years ago

Issue: corex executor doesn't work Fixes in startupcmd:

create a setter for self.pid property to be used when starting cmd using corex
2 process are running so when stopping they both need to be killed therefore in addition to killing the cmd that runs the script with the cmd_start, the actuall process of the cmd_start is also killed

Dinaamagdy commented 4 years ago

stop with background doesn't work x=j.servers.startupcmd.get("hobaddd") x.executor="background" x.cmd_start = "python -m SimpleHTTPServer" x.stop() Check that server stopped successfully but server not stopeed and status of startupcmd be stopping

Dinaamagdy commented 4 years ago

same for foreground stop not working

ranatrk commented 4 years ago

same for foreground stop not working

It is working with 2 scenarios: 1 - manually entering tmux session and starting the process in foreground and stopping it from outside tmux 2 - starting and stopping the process in foreground without tmux being involved

ranatrk commented 4 years ago

Due to the following change https://github.com/threefoldtech/jumpscaleX_core/commit/e6895989e4cbb29a58fb8ef8de36f96f39d50653#diff-03f28845f874457487974ea8fecb3dddL526 , the fixes were reverted and the following error arises again. Will look into it.

corex executor doesn't work:

startupcmd =j.servers.startupcmd.get("") j.servers.corex.default.start() corex = j.servers.corex.default.client startupcmd.executor = "corex" startupcmd.corex_client_name = corex.name startupcmd.cmd_start="python3 -m http.server" startupcmd.start()

--TRACEBACK------------------
<stdin> in <module>
    1    
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in start
    500    if not reset and self.is_running() == True:
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in is_running
    336    return self._error_raise("cannot check running don't have pid or corex_id")
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in _error_raise
    90    raise j.exceptions.Base(msg)
-----------------------------
Tue 24 08:55:35 tupcmd/StartupCMD.py -  90 - _error_raise                       : EXCEPTION: 
    error in startupcmd :ee
    cannot check running don't have pid or corex_id
None```

hossnys commented 4 years ago

start and stop working fine in foreground

3BOTDEVEL:3bot:~: tmux new-session -d -s session 'kosmos -p "j.servers.startupcmd.get().start()"'
3BOTDEVEL:3bot:~: ss -ntulp
NetidState  Recv-Q Send-Q  Local Address:Port Peer Address:Port                                      
tcp  LISTEN 0      5             0.0.0.0:8000      0.0.0.0:*    users:(("python3",pid=1793,fd=3))    
tcp  LISTEN 0      128           0.0.0.0:6379      0.0.0.0:*    users:(("redis-server",pid=68,fd=7)) 
tcp  LISTEN 0      128           0.0.0.0:22        0.0.0.0:*    users:(("sshd",pid=24,fd=3))         
tcp  LISTEN 0      128              [::]:6379         [::]:*    users:(("redis-server",pid=68,fd=6)) 
tcp  LISTEN 0      128              [::]:22           [::]:*    users:(("sshd",pid=24,fd=4))         
3BOTDEVEL:3bot:~: tmux new-session -d -s session1 'kosmos -p "j.servers.startupcmd.get().stop()"'
3BOTDEVEL:3bot:~: ss -ntulp
NetidState  Recv-Q Send-Q  Local Address:Port Peer Address:Port                                      
tcp  LISTEN 0      128           0.0.0.0:6379      0.0.0.0:*    users:(("redis-server",pid=68,fd=7)) 
tcp  LISTEN 0      128           0.0.0.0:22        0.0.0.0:*    users:(("sshd",pid=24,fd=3))         
tcp  LISTEN 0      128              [::]:6379         [::]:*    users:(("redis-server",pid=68,fd=6)) 
tcp  LISTEN 0      128              [::]:22           [::]:*    users:(("sshd",pid=24,fd=4))         
3BOTDEVEL:3bot:~:

start and stop in background is working fine also:

JSX> x=j.servers.startupcmd.get("test")             
JSX> x.executor="background"                           
JSX> x.cmd_start = "python -m SimpleHTTPServer"           
JSX> x.stop()                                          
True

but there is an issue with corex executor when you but these commands as one step the status of startupcmd is updated successfully from NOTFOUND to RUNNING and it starts without any errors.

JSX> j.servers.corex.default.start()                                                                 
JSX> corex = j.servers.corex.default.client                                                          
JSX> startupcmd = j.servers.startupcmd.get(name="startup_test",executor="corex",corex_client_name=corex.name,cmd_start="python3 -m http.server")                                                     
JSX> startupcmd.start()                                                                              
JSX> startupcmd.stop()                                                                               
True

and when you start it step by step you got an error because of failing in getting startupmd status

JSX> startupcmd =j.servers.startupcmd.get("")                                                        
JSX> j.servers.corex.default.start()                                                                 
JSX> corex = j.servers.corex.default.client 
JSX>  startupcmd.executor = "corex"                                                                   
JSX> startupcmd.corex_client_name = corex.name                                                       
JSX> startupcmd.cmd_start="python3 -m http.server"                                                   
JSX> startupcmd.start()                                                                              
Wed 20 14:55:00 tupcmd/StartupCMD.py - 544 - start                              : EXCEPTION: 
    RuntimeError()
--TRACEBACK------------------
<stdin> in <module>
    1    
/sandbox/lib/jumpscale/Jumpscale/servers/startupcmd/StartupCMD.py in start
    544    raise RuntimeError()
-----------------------------

Dinaamagdy commented 4 years ago

now all start options (corex, background,foreground ) works successfully but all .stop() don't work it return true but server keep runnning

Hamdy commented 4 years ago

@Dinaamagdy can u please update all branches to unstable and check again stop() should be working

hossnys commented 4 years ago

verified on branch unstable

Screenshot from 2020-02-12 15-01-16

Screenshot from 2020-02-12 14-56-35

threefoldtecharchive / jumpscaleX_core