#!/bin/bash

# Stop existing emhttp
kill `ps aux | grep "/usr/local/sbin/emhtt[p]" | awk '{print $2}'`

# Restart emhttp
echo "Restarting web interface"
/usr/local/sbin/emhttp  & disown

sleep 1;

# Output process list to visually check if running
ps aux | head -n 1
ps aux | grep "/usr/local/sbin/emhtt[p]";

I tried restarting emhttp with the above script (it was still running, but not responding).

 

I then received the following syslog error.

 

Jun 27 18:17:47 pius-server emhttp: unRAID System Management Utility version 6.3.5
Jun 27 18:17:47 pius-server emhttp: Copyright (C) 2005-2017, Lime Technology, Inc.
Jun 27 18:17:47 pius-server emhttp: shcmd (2): mkdir -p /boot/config/plugins/dynamix/users
Jun 27 18:17:47 pius-server emhttp: shcmd (3): mkdir -p /boot/config/shares
Jun 27 18:17:47 pius-server emhttp: shcmd (4): mkdir -p /tmp/emhttp
Jun 27 18:17:47 pius-server emhttp: shcmd (5): mkdir -p /var/local/emhttp/smart
Jun 27 18:17:47 pius-server emhttp: err: device_by_name: device_by_name: (sdh) not found

 

Edited June 27, 20179 yr by benyanke

Unless things have changed (doubt it), emhttp has never been designed to be restarted

So what do you do if everything is running normally and emhttp hangs? It happens more than occasionally. This is linux we're talking about here. It's silly to bring down a system to restart a daemon.

Edited June 27, 20179 yr by benyanke

It is what it is.  You reboot.

 

6.4 utilizes nginx as the engine, and is far harder for the UI to actually crash.

Bringing down the whole system (including containsers and VMs) when only a secondary management interface is really unfortunate... I hope this can be remedied at some point. The whole reason I went for a linux based system is to avoid things like this.

 

Is there any timeline or plans to set up the service properly so it can be restarted like a normal linux service?

Edited June 28, 20179 yr by benyanke

Update: even upon restarting, the web gui won't respond. emhttp is a running process still, however. What debugging info do I need?

Edited June 28, 20179 yr by benyanke

'Tools'    >>>>  'Diagnostics'.   Then attach the entire file to your next post.

But I can't access the web gui....

diagnostics from the command prompt. Would also be good to post as a feature req as LT doesn't read every thread

Sent from my LG-D852 using Tapatalk

SORRY!    I assume that you can log in via either the console or a Telnet session.   Then type  diagnostics  on the command line.  that will write the diagnostics file to the logs directory/folder on your Flash Drive.  

3 hours ago, benyanke said:

Bringing down the whole system (including containsers and VMs) when only a secondary management interface is really unfortunate... I hope this can be remedied at some point. The whole reason I went for a linux based system is to avoid things like this.

 

Is there any timeline or plans to set up the service properly so it can be restarted like a normal linux service?

unRAID is a not a fully open Linux implementation.  It is an appliance style implementation with an embedded OS.  You get many of the benefits of being on an open Linux platform, but not all.  That said, I think your request makes a lot of sense and hope you follow Squid's suggestion to post it as a feature request.

Since the introduction of unRAID 5 it is not possible to restart the emhttp process. I believe that was a deliberate design choice of LT, but as suggested you can make a feature request and give good (convincing) arguments why it needs to change.