|
Disk get corrupt way too often for a normal use
Hello,
I'm running a vps under debian and every few weeks my disk suddenly switch to a read-only mode. The step of resolving this have become a habit. I reboot, the disk can't be booted but the prompt is asking for my root password, I login, I run e2fsck, I remount the disk and reboot. So, what's the problem?
The problem is that, my server isn't even one year old and that it is the fourth time that this problem occur. I suppose one of my services running must be causing this problem but I failed to find which one. Here's the complete list of deamon runnings (content of /etc/init.d/) :
acpid kmod redis-server
apache2 motd rmnologin
atd mountall-bootclean.sh rpcbind
avahi-daemon mountall.sh rsync
bootlogs mountdevsubfs.sh rsyslog
bootmisc.sh mountkernfs.sh saslauthd
checkfs.sh mountnfs-bootclean.sh sendsigs
checkroot-bootclean.sh mountnfs.sh single
checkroot.sh mtab.sh skeleton
console-setup mumble-server snmpd
cron mysql spamassassin
danted networking ssh
dbus nfs-common udev
dovecot ntp udev-mtab
exim4 postgresql umountfs
halt procps umountnfs.sh
hostname.sh rc umountroot
hwclock.sh rc.local unattended-upgrades
kbd rcS urandom
keyboard-setup README x11-common
killprocs reboot
And inspircd run manually on startup.
So, two question. Do you have any idea of what program could corrupt the file system? If not, how can I find out ? And, second question, how can I automate the recovery procedure each time the disk switch to read-only mode? The latter I find out there is a problem, the longer my website is unusable.
This isn't my first sys admin experience but it's the first time I host things that people actually use and theses problems are starting to make them really unhappy of my services. :/
|
