I had a web server that got overloaded. I attempted to stop apache but it just hung so I kill -9 the apache pid.
When trying to run
sudo /etc/init.d.apache2 start
or any derivative thereof, nothing happens. The command doesn't return. There is no error in log or at command line. I just get Starting web server: apache2 and the cursor keeps blinking. I've waited 30 minutes and still nothing. The server has 10G of memory free.
I running Debian wheezy.
Start by looking at the log files. That should contain the reason.
If there is nothing in the log files, troubleshoot using strace shown here : https://serverfault.com/questions/150249/best-way-to-troubleshoot-apache-not-starting
strace showed it was hanging on an nfs4 mountpoint that was not functioning properly.
Related
I have installed the Hue on the Linux whixh is an instance from Azure. I have made all the required changes in ambari and hue.ini conf file. And when I run the supervisor job, it runs fine
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
hue 83933 sshuser 3u IPv4 15707246 0t0 TCP *:8000 (LISTEN)
But when I try to access the wb hue, I don't get any page loaded. It shows refused to connect.
Tried deleting caches and reset up was done.
I am using hue 4.7 version and I don't find any issues in error.log file. Yet, I don't see any data in access.log file. Could you please help me?
Do you have
http_host=0.0.0.0
in the hue.ini?
#Ruthikajawar here is a working hue.ini for ambari
https://github.com/steven-dfheinz/HDP3-Hue-Service/blob/Hue.4.6.0/configuration/live.hue.ini
I have noticed that sometimes, after initial install, it takes 1 or 2 restarts to get the WEBUI to work. I have also noticed sometimes, after a restart, it takes quite a few moments before the WEBUI starts to respond.
Give it some time after restart and check the WEBUI. If you still are not getting it to answer you need to check /var/log/hue/error.log as it should be very specific with errors causing the WEBUI to fail on startup.
Below is the BHd.service file; this is the first time I've done this, and everything is finally working perfectly except that when I stop the service, it always without fail stops httpd (and likely other services listed here). I sincerely have looked everywhere for the answer. I don't think it should be doing this.
Log story short, I need smb, nfs, httpd and mariadb to be running before this unit start. I do not want them to be stopped after the unit is stopped or reloaded; for now, firewalld must be off. I honestly can't tell what line is affecting the systemctl stop command, everything I read indicates that Requires and After only affect systemctl start.
[Unit]
Description=BH
Documentation=somewhere.com
Wants=smb.service nfs.service
After=httpd.service mariadb.service
Requires=httpd.service mariadb.service
Conflicts=firewalld.service
[Service]
Type=forking
PIDFile=/var/run/BHd.pid
ExecStart=/usr/bin/python /var/www/html/pythonscripts/BHd.py start
ExecStop=/usr/bin/python /var/www/html/pythonscripts/BHd.py stop
[Install]
Alias=BHd
WantedBy=smb.service nfs.service
RequiredBy=httpd.service mariadb.service
EDIT: More info
[root#BHDEMO ~]# systemctl -l | grep BH
BHd.service loaded failed failed Description
[root#BHDEMO ~]#
Short Answer: systemctl disable BHd
Long Answer:
Well, I feel stupid... I'm sure some of you will think "this guy is writing a 40k line service but didn't know that!"... yes, that's how us computer engineers roll, lol.
The Solution:
After editing this file over and over in an attempt to get it working, I never tried systemctl disable BHd... that did it... I had a few lines in the Install section early, WantedBy and RequiredBy as seen above. I realized about 5 hours ago that this was not correct and deleted it. Unfortunately, just doing another enable is not sufficient, as those lines created directories that needed to be uninstalled with disable, then reinstalled with enabled without the added files and folders. Almost sounds like a linux bug that should be fixed, imagine how many like me over the years wasted so much time.
I've been having a problem with my (development) site hanging when I try to login using LDAP credentials. Using xdebug, I was able to pinpoint the hang to a specific line of code, which is a call to ldap_bind(...). After days of trying to understand why it is hanging, one of my debugging techniques was to disable mod_ldap, and then just try to get any error to show up in apache log (after this has started, apache error logs haven't been recording any errors while performing this http request; It 'hangs' too).
What I did
I disabled the module (sudo a2dismod mod_ldap), restarted the server (sudo service apache2 restart), and confirmed that the module isn't enabled (apache2ctl -M does not show mod_ldap)
The Problem
It still hangs when it reaches ldap_bind(), however, it shouldn't even be reaching that point because without mod_ldap, my code shouldn't even be successfully calling ldap_connect() (right?) which is returning (resource) resource id='5' type='ldap link' (meaning the call was successful). I'm expecting a NoMethod or Function error.
Why can php make a call to a module that isn't enabled, and how do I stop that behavior?
Version
Ubuntu 14.04
Apache 2.4.7
PHP 5.5.9-1ubuntu4.5
Well it turns out, I had a novice understanding of what was happening. Never occurred to me that ldap_connect() as a php call had nothing to do with mod_ldap or any apache module. Thus, disabling/enabling apache mods wasn't doing anything because... it had nothing to do with apache mods.
So... I was able to disable ldap php extension with
sudo php5dismod ldap
and able to check that it was disabled with
php -m
And sure enough, after restarting the server, running my script caused a No function error for ldap_connect() (just as I was expecting/hoping). Now I just need to figure out why ldap_bind() hangs...
I'm hoping someone out there is used to monit and can help me.
Im running a home data server, with Ubuntu 13.10.
I have CGminer setup to start when the PC boots, from a bash script of my own creation. It contains a few tweaks and setting that need running before it gets going.
But if for some reason my interweb goes down...cgminer will close after a small amount of time. Now, if im asleep, and it closes. That valuable mining time, and a waste of the electric. So I'm looking into monit as a way of fixing that.
Im hoping to be able to have monit (or something similar, doesnt have to be monit) Start CGMiner from my script, check every so often that CGminer is still running, and if not, restart it from my script.
I just cant get my head around the config file for monit...Help would be awesome
Yes, you can achieve that with monit. You only need that your start script writes pid into pidfile:
check process xyz with pidfile /var/run/xyz.pid
start = "/bin/xyz start"
stop = "/bin/xyz stop"
i want to set a script which will restart the server automatically.
And i write the following code to setenv.sh file
JAVA_OPTS="$JAVA_OPTS -XX:OnOutOfMemoryError=/usr/local/apache-tomcat-5.5.30/bin/shutdown.sh;/usr/local/apache-tomcat-5.5.30/bin/startup.sh;"
It not working properly.
I am using tomcat 5.
To answer your question - I don't think the error handler can run several commands. If you want to do this, then write a small restart script and run your custom script.
However, I would think about this again. Not every OutOfMemoryError means the server should be restarted automatically. You'd better to get a notification and then profile the server to get the cause of the OOME, so you can get rid of it.
P.S
Any reason to use the old tomcat 5.5? tomcat 7 is quite mature.
Use CATALINA_OPTS instead
CATALINA_OPTS="${CATALINA_OPTS} -XX:OnOutOfMemoryError=\"/bin/sleep 30;/bin/kill %p; /bin/sleep 60; /bin/kill -9 %p\""