Server :: Disk Filling Up - How To Track Down Rogue Files
Feb 5, 2010
I'm running an OpenSuse 11.2 box at home which updates a mysql database 4 times a day then posts 34Mb to my website. It's exactly the same amount of data each time, and my scripts*TRUNCATEs then rewrites the database with the latest data - so the database size remains the same.
There's a problem (I think with the script) however which means that everytime the script runs, approximately 34Mb of space on my hard disk is mysteriously taken up. I'll have to get that script fixed...
I can't, however find the files which are eating up my diskspace at the rate of 140Mb per day. I've done various searches (mainly with Dolphin) including hidden files immediately after running my flaky script and looking for any files created/modified in the previous few minutes when the 34Mb has disappeared.
There are LOTS of files in /proc (which I don't think is actually on my HD right?) and also in /var. There's nothing much in /tmp (on a separate partition) or any log files that I can see. The box has been running this script daily for the last 6 weeks so I'm hoping there's a load of files somewhere I can get rid of, then fix my script.
For a while my root partition has been filling up for no apparent reason. I I have been deleting things to find out it fills up again in a mater of days. To make it more 'interesting', there is a disparity between what I get from df and what the du command is telling me. After dismounting the other file systems and turning off applications, this is what i get:
du claims that I'm using 29G on that partition, which sounds about right (this is my OS and basic /home partition, everything else is elsewhere). df on the other hand is telling me here that out of 69G, 64G are in use with only 54M left.
I have a lot of data on a nfs store external to an Ubuntu machine. I've had problems with the storage becoming full and have spent a very long time deleting files to get absolutely nowhere! Over 25GB of deletions on the 80GB disk has only cleared 1.5GB of space...
You can see below the hidden .snapshot directory, the disk usage for the nfs-store is coming out at 97GB when the disk is only 80GB in size, and the two directories I want on the disk are about 22GB in size.The .snapshot directory appears to have been made on Friday, I'd like to know if I can:
1) find out what command ran to create/update it
2) re-run the command to update the directory or remove the directory
Code:
ideasadmin@ideasadmin-desktop:/nfs-store$ ls -la total 20 drwxr-xr-x 5 500 500 4096 2010-07-16 17:35 .
Noticed that one of my partitions on openSuse 11.1, mounted as /usr, filled up suddenly with over 20 GB of something. Tracked it down to /usr/bin, in which the X11 directory contains 2353 items and another X11 directory, recursively at least 11 times that I've expanded without reaching the last X11 directory. The newest files in each directory are dated 11/24/10.
Any idea of what's happening? How to stop it before it fills the disk?
I have a directory /var/log/data its about 80 GB,It filling up quit rapidly.I don't have much space left in the system them So i will attaching another External HDD.My question is that i need to mount /var/log/data to new HDD.So i have old data and pulse new coming up.I don't want to copy data from /var/log/data then mount new HDD to /var/log/data you know what i am taking about is there a simple way like linking or any other.
I would expect this behavior if "domain.com and anotherdomain.com" wasn't a domain that I hosted. But this is a valid domain that this server should be answering for. In my named.conf I do have the
Quote:
allow-query { any; };
option on every zone. This is my slave server and I have the primary shut off so I can test this slave server. FYI: So far queriers still seem to be working. The pages for the sites are still coming up via the internet.
In the last week email delivery had slowed down to taking approximately 24 hours and investigation has revealed that there were 45000 emails in the mailqueue. I flushed the mailqueue yesterday but mail delivery is slowing down again because emails are again being queued by the thousand.
The mails are mainly spam sent to ficticious users at a genuine domain hosted on the server.
My setup is running Courier IMAP with Postfix, Spassassin is installed along with Amavisd-new but I just don't understand why this has suddenly become a problem based on my setup not having changed for the last 6 months.
Should these spam emails be being discarded by spamassassin before reaching the mail queue? Are my discard score settings too high (6.9)?
I am using Fedora 13 64bit. My eclipse program starts a server on port 9050 but I shutdown the program, eclipse fails to properly shutdown the server. Restarting eclipse has no effect.
How can I close or unbind a port forcibly from command line?
I know rebooting or log off and back on will clear the ports but this much slower process for me. I am looking for forcibly closing the port in command line.
I work for a company that makes portable devices running Linux and I was recently asked to make the underlying file system read-only for "security" purposes. Since the distribution is based on LinuxFromScratch, I know that very little writing happens at run time. So, even if the device runs on a usb flash device, I doubt that putting the root file system RO will be that beneficial. I am actually more concerned about a process actually breaking because it cannot open a file in RW mode than a process going rogue and filling the root file system with log files, etc. I'd really like to ear what kind of advantages disadvantages there really is with read-only file-systems.
I am trying to find a best tool to track configuration files changes. I did find some information about osec and mactime, but, it seems, that they are not included in fedora/rpmfusion package databases. is there any tool that can be installed as a package?
I am writing a client chat server program in java. The chat server is already written. I created the interface and I connected to the chat server. The server keeps sending protocol messages to whom ever is connected, also it keeps sending specific protocol every 10 sec to check if user is still online.My problem is when I connect from my interface to the chat server, the program gets stuck, I can not press any other button because after connection my program keeps receiving commands from server(I have a method to keep reading after connection). How can I keep track of messages sent from server and not get stuck i.e press other buttons or type in the interface?
I need to trace/track email activity in my ubuntu server, normal i use this command tail -f /var/log/mail/mail.log, more or less. But the problem i am facing is that it take a long time to go through the entire log as my server recieves thousands of emails per day. is the a code that i can use to search email in the log, for instance need to find perfectpol@mycompany.com in the mail log.
Secondly; if there is an email that is blocking others on the queue how can it be remove or forced to go the mailbox. or to view its contents in order to decide whether to remove it or not.
Had a problem that are said not arrival or that are delivered as late as a day and sometime 6hours.
I am having problem with the root disk / fill up. So now that i have the disk information provided from SAN team Id 0:3, how would i do to add this device to / to increase space on Redhat Enterprise server 4?
I'm booting to Kali 2.0 live from USB and wanted to add persistence, but I can't get OpenVAS setup. The setup script runs and eventually fails due to no more disk space. Here's my df -h output:
Here's gparted:
When the setup runs it fills up root (/) which is only 872mb. This is a 16gb USB so I'm wondering if there's a way to allocate some of the 11gb of unallocated space to root? I couldn't tell how to do this with gparted, would I need to build a custom Kali iso or something with different partitioning?
I have a pre-printed form that I need to fill in. Is it possible to scan it, fill it in on screen, and then put the original form in the printer and get things to print out on that original form? I know that I can scan the form and fill it in on screen and print out on a blank piece of paper, but I need to use the original form
My /tmp directory is being filled up with root-tmp.####. I suspect they are being created by bastille-tmpdir-defense.sh, but they do not seem to get removed.
I am relatively new to ubuntu/linux and although I have made some good progress on my server I'm struggling with a few points. I am sure what I'm about to ask has been covered in some other thread/guide but I just cant pick out the missing piece hence my direct question;I have 2 computers; a server and mediacenter.on the server I have installed nfs:
Code: sudo apt-get install nfs-kernel-server nfs-common portmap then
Just today I started getting notices about lack of disk space on my system. After much digging I found that .xsession-errors and .xsession-errors.old were taking nearly 70GB of space combined. The primary message I'm getting over and over again is: SSL_Write: I/O Error I have been unable to figure out what's causing this error.
Some of my cronjobs are filling up files in my / directory. How do I make this stop? One of my cron jobs uses wget:[URL].. The bexcb0.php file writes a file and then echos a result if it is sucsesfull. These echo results are being put into bexcb0.php files in my /root folder and are piling up.
My / folder is filling up with files bexcb0.php etc bexcb0.php bexcb0.php.1 bexcb0.php.2 bexcb0.php.3 bexcb0.php.4 bexcb0.php.5 etc How do I make this stop? If I just remove the echo will they stop writing to the / folder?
I am using my ubuntu server as my home router.Everything is working as expected with one exception.y DSL modem is a POS and every now and again it looses connection to the router. Sometimes it needs to be reset and sometimes it does not.Either way, when this happens my ubuntu server needs to reacquire an IP from my ISP. If it screws up when I'm at home it's no big deal, but if it happens when I'm not around my housemates have taken to hitting the reset switch on the server. I'm not a big fan of this so I wrote a script to ping my ISP's gateway. If it's unavailable it bounces the eth0 interface and tries to get an IP. I am running this script every couple of minutes in a cron job. Now I'm getting syslog entries like
Code: Aug 9 20:31:01 portal CRON[9602]: (root) CMD (/opt/ChkAndFixNetwork.sh) every few minutes. This is annoying and makes the logs useless for troubleshooting. I
I have some ideas about writing a small game in terminal ( just for fun ) using ncurses library. I want to use some kind of menus (in Midnight Commander's style), but there are some problems with rendering windows, that I don't understand. I create a window with newwin(), assign a color pair to it calling wattron() (for example, I want to fill a window with blue background), and then I call my own function wnd_fill() :
When building 8.04 servers, I reconfigure snmpd's logging options to prevent copious low priority messages being logged whenever our network management workstation polls them. I edit /etc/default/snmpd and change line 11 from:
