hda1/dma issue

[I Love Jesus]

In WHM's server status I get:

Disk hda1 (/boot) DMA+ 82% (flashing yellow button status)

Yellow is like a warning sign in the status area so why is it flashing yellow for this boot partition, and how do I fix this?

[dezignguy]

probably because it's over 80% full...

You should probably uninstall a few old kernels... keep at least two new ones though, if you should have to revert back for whatever reason.

[I Love Jesus]

Could you please be kind enough to explain to me how to remove old kernels? Plus, this server was freshly re-imaged only 2-3 weeks ago. I'm not quite sure how fast new kernels are updated by the server user but I cannot imagine that much, that quick.

[EdRooney]

ls /boot

Let us know whats in their.

Look for kernel images with kernels you don't need anymore and remove those, then edit you grub or lilo conf file accordingly. I recommend keeping at least 2 different kernel versions. Also, I like to make /boot 1gig, harddrive space is cheap, you never know what size the next kernel will be, or a kernel 5 years from now.

Hope this helps.

[dgbaker]

ls /boot

man ls

I highly recommend familiarizing yourself with the ls command.

Nice answer Abe

You give someone an answer like that? You're the one constantly asking similar type questions. I suggest you take your own advise sometime.

[EdRooney]

Is
ls /boot
not the place to start?

How am I suppose to know what he has in there.

[I Love Jesus]

root@server [~]# ls
./               bfd-0.4/              firstboot_log         public_ftp/
../              bfd-current.tar.gz.1  .forward              public_html/
.accesshash      .cpan/                .gconfd/              .rnd
anaconda-ks.cfg  cpanel3-skel/         Install_DigiChat.bin  .spamassassin/
apf-0.9.4-7/     cpanel_install.log    install.log           .tcshrc
.bash_history    .cpcpan/              install.log.syslog    urchin-add.php*
.bash_logout     .cpmysqlrpm/          mail/                 urchin-remove.php*
.bash_profile    .cshrc                .my.cnf*              .Xresources
.bashrc          download.php?did\=54  .pgpass

root@server [~]# ls /boot
./                              quota.user*
../                             System.map@
aquota.user*                    System.map-2.4.21-20.0.1.EL
config-2.4.21-20.0.1.EL         System.map-2.4.21-20.0.1.ELsmp
config-2.4.21-20.0.1.ELsmp      System.map-2.4.21-20.EL
config-2.4.21-20.EL             System.map-2.4.21-20.ELsmp
config-2.4.21-20.ELsmp          System.map-2.4.21-4.EL
config-2.4.21-4.EL              System.map-2.4.21-4.ELsmp
config-2.4.21-4.ELsmp           vmlinux-2.4.21-20.0.1.EL*
grub/                           vmlinux-2.4.21-20.0.1.ELsmp*
initrd-2.4.21-20.0.1.EL.img     vmlinux-2.4.21-20.EL*
initrd-2.4.21-20.0.1.ELsmp.img  vmlinux-2.4.21-20.ELsmp*
initrd-2.4.21-20.EL.img         vmlinux-2.4.21-4.EL*
initrd-2.4.21-20.ELsmp.img      vmlinux-2.4.21-4.ELsmp*
initrd-2.4.21-4.EL.img          vmlinuz-2.4.21-20.0.1.EL
initrd-2.4.21-4.ELsmp.img       vmlinuz-2.4.21-20.0.1.ELsmp
kernel.h                        vmlinuz-2.4.21-20.EL
lost+found/                     vmlinuz-2.4.21-20.ELsmp
message                         vmlinuz-2.4.21-4.EL
message.ja                      vmlinuz-2.4.21-4.ELsmp
root@server [~]#

[EdRooney]

80% isnt bad necessarily, I would remove these 2
System.map-2.4.21-20.EL
System.map-2.4.21-20.ELsmp

[casey]

You can do:

rpm -qa | grep kernel

and then you can see what kernels you have. To delete the ones you don't need, run

rpm -e name_of_the_kernel

It will remove all the unnecessary stuff for you.

[I Love Jesus]

Forgive me for asking, but how do I know which ones I need and do not need? I've never deleted kernels so I just want to be careful.

[EdRooney]

If its working ok with the latest version of the kernel, keep that one, start by deleting the oldest, then second oldest, etc until you have enough free space on /boot.

[dezignguy]

initrd-2.4.21-4.EL.img
initrd-2.4.21-4.ELsmp.img

Those are old ones... I'd remove those... there's no reason you'd want to go back to those. Leave the 2.4.21-20 versions as those are the two most recent kernels.

Clarifying my last post, 80% isn't necessarily bad... since the /boot partition isn't going to be growing as the server is used and cause problems when it hits 100%, like /home. So it's not too much of a worry as long as you remember to remove old kernels more frequently, as you update to new kernels. Since you seem to have a rather small /boot partition. I've got 9 old kernels (some are doubled up with the SMP version) still installed and am barely over 50% usage.

[chirpy]

And please, if you are usng a RedHat OS, you really should follow casey's advice and use rpm to remove any package installed on your system, especially old kernels. Not doing so could render your server unbootable thanks to the simplest of typos. Unless you are a competent system admiinistrator you shouldn't touch anything in /boot directly.

[I Love Jesus]

root@server [/boot]# rpm -qa | grep kernel
rpmdb: Program version 4.2 doesn't match environment version
error: db4 error(22) from dbenv->open: Invalid argument
error: cannot open Packages index using db3 - Invalid argument (22)
error: cannot open Packages database in /var/lib/rpm


???

[haze]

Try rebuilding the rpm db: rpm -vv --rebuilddb

uname -a will show you what kernel you are currently running. You can not remove a kernel you are currently live on, keep that in mind.

Before you remove ANY kernel, ensure you are on the latest, which is now: 2.4.21-27.EL
Its a good rule of thumb to keep the last 1 - 2 kernels you know have worked, just incase the new kernel buggers up, or has problems in the future you can then revert back to a known working kernel.

Is this a hyperthreaded or dual processor machine ? If not, get rid of ALL of the SMP kernel's, if so, get rid of ALL of the NON SMP kernels.

When you are complete, I suggest your audit your lilo.conf or grub.conf ( depending on which is in use on your system ) to ensure they are correct and your newest or current kernel is selected.