Software :: Restart Apache2 Process Every 10 Mins To Clear /tmp Directory?
Aug 31, 2009
I have been working on this for a few hours now and still haven't been able to track down whats going on. I just switched servers and put a fresh debian install on with apache2 + custom suexec + php5(fcgi). now when i upload a file to the site it wont upload, so i tracked down the issue. its the /tmp directory that root tmp. its at 100% so i restart the /etc/init.d/apache2 process and the /tmp directory is automagically cleared to 20% used and i can upload again. so i figure ok its got to be a setting. i looked in all the conf files conf.d the php5 conf files and no where do i have a /tmp directory set. so i think OK I can build a cron to clear the tmp directory every 10 mins till i find the issue, but alas, when i go into the tmp directory There are NO files to delete!!! wtf. holy crap. and all that goodness, this is crazy, how can you have a tmp directory that is not being written to, from all i can see, so i think ok it might be a hard link, i checked, in all root and sub root directories no hard links to tmp, no soft links to temp either. and if i sit in my /tmp directory and do a df -h after some uploads its going from 20% so 23% and up, but i do ls -alh and nothing, then ls -al and watch bytes, nothing , so this is very strange. has anyone encountered something like this? is it possible that this is a hard link thats hidden somewhere or some other extreme function i have never heard of?
it seemed like the most accurate place. Also apologies for any inaccurate terminology, I'm a bit new at this. Running Apache2 on top of Debian 5.0
Anyway, I have irssi set to output logs to a folder accessible by my web server. User permissions are all set up, so it writes to the folder just fine, but when I access the server index in a web browser (i.e., page that says "index of /[directory] at the top) I cannot see the the directory or the logs that irssi is making. I can ssh in and see the folders and files in the terminal, so they are being created. How can I set it so these are viewable through the web server? I tried restarting the server, no effect.
I am having some problems after installing apache2-mpm-itk on my Debian 5.4 server. From my understanding apache2-mpm-itk will run child processes as root until a http request is received. It then suid's to the username in the VirtualHost directive. This is all working fine, as I can see a few apache processes started by a specific user.
However, the problem is when I try to restart apache. It will kill all the processes running by root but wont kill any of the processes that have been suid'ed to another username.
When i attempt to restart I get this error:
Code: charged:/etc/apache2# /etc/init.d/apache2 restart Restarting web server: apache2 ... waiting (98)Address already in use: make_sock: could not bind to address [::]:80 (98)Address already in use: make_sock: could not bind to address 0.0.0.0:80
[Code]....
If I do a "ps aux | grep apache" before and after I see all the root processes killed, but processes running as specified users from VritualHost directive remaining.
Now I am not an extreme linux power user, but I know a bit.
My system is old. Not debian.How do I clear out the mqueue directory? I don't want to delete it....just clear it out of the email that is clogging up my var partition. This old command doesn't seem to work anymore:find /var/spool/mqueue -exec rm'{}' ;Anybody know of the command to clear my mqueue.
I currently have a massive problem with my main webserver: one of the SCSI drives where I keep my htdocs seems to have failed (sd 0:0:0:0: [sda] Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK,SUGGEST_OK end_request: I/O error, dev sda, sector 7797601).
Since I have no backup (dumb!dumb!dumb!) of it, I tried to copy what I could on the fly, but that failed miserably. The next idea was to umount the volume and try remounting it R-O, but I have a bunch of apache2 processes that are apparently reading/writing on it and I can't seem to find any way of terminating them. I tried via init script, killall and kill -9 but they just don't respond. Until I can terminate those processes, I won't be able to unmount the drive and try saving what's left of the data.Oh and the cherry on top of the cake: the server is 1100 miles away so I can't even access it physically..Is there any way I didn't think of to get rid of those apache processes?
I'm using debian 5 x64 with xfce.Is there a way to configure (start/stop/restart) services (especially Apache2, mySQL and PHP) using a graphical or cli tool? I tried to use sysv-rc, sysv-rc-conf, rcconf and rc-conf in the terminal but Bash didn't find them (Although Synaptic show that sysv-rc is installed).
I am trying to access log file which located in /etc/log/apache2. I could get into the directory using `su`. I was able to run ls command under the directory and everything was file. I could run a command,
ls -d /var/log/apache2/*
However after I switched to my account, I got an error. sudo ls -d /var/log/apache2/* ls: cannot access /var/log/apache2/*: No such file or directory
I want to use this command in a bash script to get a list of log files. Should I write the script as root and run it as root?
I'm running the turnkey linux version of joomla and I need to use PHP's mail functionality to send a notification email from a page (separate from joomla content) . PHP's native stuff is rather plain and I find myself needing the functionality of PEAR. So I've installed PEAR via aptitude, but it has installed to /usr/share/php.
how do I include these files from php scripts? Firstly, the files are owned by root, the www-data user doesn't have access. It also appears that the permissions on the files may need to be modified. Some php files don't have the execute flag (shouldn't they)? I recognize the point of having a centralized location for updates and sharing, so I don't want to just copy the files to my web directory (/var/www).
I didn't know exactly where to put this, but here it goes:For the record, I have Apache2 on Ubuntu.OK so I have a "svn" subdomain for SVN repositories. The repositories are actual separate repositories, not a big one for all. So I'm using SVNPath not SVNParentPath (please don't suggest this as a possible fix).
The problem is I want to be able to list the repositories (or even put a blank page for that matter) at svn.mydomain.com. All repositories have a svn.mydomain.com/myrepository path with a VirtualHost for each. So when I create another VirtualHost just for the / path, it overrides all the SVN VirtualHosts, as it covers / and all sub-directories.
I would like a way to set that VirtualHost to affect just the root directory, without the sub-directories. Or to make the SVN VirtualHosts override it. Either way, is this possible? What approach should I take? Thanks!
i have a web site and i need to forward all requests to the same page. Basically i would like to show a maintenance page while we are working with our database behind the application and in the meantime i would like to redirect all url in the root directory as following :
http:[url].....
http:[url]....
i tried to perform the action using mod_rewrite in this manner :
Code: RewriteEngine On RewriteRule . http://www.mysite.com [L][code].....
The webpage at http:[url]....has resulted in too many redirects. Clearing your cookies for this site or allowing third-party cookies may fix the problem. If not, it is possibly a server configuration issue and not a problem with your computer.
I have been searching the web and Ubuntu forums all day looking for an answer to what I initially thought would be a simple thing to do. I have set up a lamp server and started creating my network site, however was not long before I ran into an issue with fire fox not allowing me to access locally linked files that are shared on the network which works fine in IE. after several hours of trying to find a work around, I have been unsuccessful in creating a link that would work in both IE and FF. so I got this bright idea see, to just make the website include the network shared content, unsharing them of course and simply adding them to my Apache server, well the more I read the more confusing it all became. do I use simlinks or do I create a server for each location/hard drive? I really need a how to for big dummies on adding additional directories to Apache similar to adding virtual directors to IIS for windows.
I have a process that I cannot kill with kill -9 how to go about this?
[Code]...
It is an openvpn process but I cannot retsrat the service as I alreday have another openvpn service running on the server so when I do openvpn service restart, it won;t know which service to restart.
Suppose a process creates multiple shared mem segments as part of it job.Now is it always possible that on restart, that process can reattach to those previously created shared mem segments to the same virtual address spaces (returned on first time creation) ? My intention is use pointers over shared memory.So on restart I want those pointers absolute value to remain valid.
I have forgotten my administrative password and am locked out of my system. I have an HP Mini 1120NR and I am running Edubuntu software. I cannot remember which key I need to press during the restart process to bring up the system restore/recovery option.
There are two host agent process ( hprsmha -WD and RSMHA ) running on my SLES 10 server suddenly hangs which is critical for my daily scheduled backup jobs.I want to create a script which checks for the 2 process and restart it automatically.so i can put it to daily cron job.My knowledge with scripting is so poor. Following are the commands i am running manually to stop and start the process.
cd /etc/init.d ./hprmsha stop ./hprsmha_wd stop ./hprsmha_wd start ./hprsmha start
I have a java service that needs to be run in a gnu screen session so that I can -r into it to access it's interactive shell. I would like this to run at start-up. More importantly, it crashes frequently so I need the service to restart itself when it crashes. How do I go about doing this myself? My previous experience with shell scripting amounts to putting ./service in a file and marking it executable. This will be running on a FreeBSD server, but I might put it on a Linux server soon also. So any caveats on the differences would be useful information.
I haveing a big problem atm (just after to upgrade to 10.04). I have a simply and easy script that sent a halt command to every KVM virtual machine actually running into the server. this script is/and was) palced as a init-rc script in rc0/rc1 and rc6 lvls. Just after to upgrade, I have noticed that this script is executed but the system ignore the time that it require to finish (giving time to correctly stop a single virtual machine). Aftert 5 second it start sending SIG-TERM and SIGN-KILL to any currect process and then reboot/halt.
I have just done a new script using upstart and I got just the same result. What is happending? why is the waiting the init secuence for a valid/or_not termination of any init script. I am so sorry for this crap english I have never study this language so I try to do my best anytime I speak it (or try ti..).
I wanted to put 11.04 on my wife's Win 7 Toshiba laptop so thought I would live boot and see everything worked first. Well it does seem to but I didn't have time to install it then so just shut down. Well on restarting without CD in it failed to boot windows and had to let the recovery process fix it. Now I want to set it up for dual boot at least to start with so until I can get some answers I don't want to risk it. Is there some problem with win 7? Will I be able to shrink the Win7 partition to put ubuntu on? Will Win 7 have an error each time she boots it after 11.04 is installed and run?
I'm very very tired, worked all night long, and I did't sleep for hours... So I'm like a zombie now... half awake, and half asleep.
I was trying to clear a directory. Then I run the command cd to enter the directory, and then before thinking I run this dangerous command. on my Linux server: "find / -mtime +1 -exec rm {} ;"
I got a lot of:
Could this command have deleted something inside these directories? I'm afraid that the next reboot the server won't startup...
I have suse10 64 bit installed. I am setting up a svn server on it. After installation and adding the modules ,while reloading the apache2 it's throwing the error as: HTML Code: httpd2-prefork: Syntax error on line 113 of /etc/apache2/httpd.conf: Syntax error on line 31 of /etc/apache2/sysconfig.d/loadmodule.conf: Cannot load /usr/lib64/apache2/mod_dav_svn.so into server: /usr/lib64/libsvn_subr-1.so.0: undefined symbol: apr_memcache_add_server
Starting web server: apache2[Wed Dec 09 15:36:40 2009] [warn] NameVirtualHost XX.XX.XX.XXX:80 has no VirtualHosts(99)Cannot assign requested address: make_sock: could not bind to address 68.178.232.100:80 no listening sockets available, shutting down Unable to open logs failed!
I have udev rules that creates a directory for each usb-media that is plugged into my debian to automount usb-media. This directory where mount-points (the mount-directories) are created is passed through a rdp-connection with rdesktop. On our terminalserver there's a drive Z that shows all mount-points with content of usb-media.When users remove usb-media, udev unmounts usb-media and tries to delete the directory that was used to mount usb-media.If the users hasn't closed explorer showing any subfolder of usb-media, the unmount-command succeeds but the rmdir-command fails because the ressource is "busy or used".My question: how can I force to delete this mountpoint? (rmdir doesn't have force and rm -rf doesn't work)Edit: I should add, that I can't use third-party software or additional packages because the linux-machine is a thin-client with very low disk-space.
