[nSLUG] Load spikes, X bogs down. Why? Fix?
mspencer at tallships.ca
Thu Apr 7 00:31:04 ADT 2016
The problem is that, on a new setup, the system bogs down so badly
under some conditions (appareently during HD access) that the X cursor
takes seconds to respond and X window manipulations take even longer.
Here's the background:
After a HD failure, I was happily able to reinstall Linux from
original media, restore personal files from backup only a few days
old, redo the numerous tweaks for my idiosyncratic customization
including using my old xorg.conf.
The only thing that has changed (unless I've overlooked something) is
that the old drive was IDE, the new drive SATA.
The IBM P4 mobo has two plugs labeled "SATA IDE 1" and "SATA IDE 2".
The failed HD was IDE, the replacement SATA, plugged into one of those
mobo plugs. Linux sees the new drive as HDA, not SDA, so that's
something I'm unfamiliar with.
BUT: running a command like:
cp /mnt/usb-thumb-drive/700-meg-file /mnt/hda3/storage
causes X to bog down and respond very laggardly to mouse movements &
clicks. Keyboard response is similarly very slow. xload(1) shows an
enormous spike until the copy completes.
A somewhat similar effect (slightly stuttery mouse cursor response,
xload spike) is seen at some other times, e.g. when mplayer is
starting up to play a Bach piece from a HD file (although not after
playing has begun). Didn't used to do that.
I only have 500M of RAM but I've been using this box a long time
without this problem.
Is there some difference in how a SATA drive versus an IDE drive is
managed that's causing this? Or might there be some similar overhead
because of the SATA->IDE emulation?
Or any other suggestion/explanation about this?
Other system details below,
Kernel: Linux 188.8.131.52
CPU: i686 (IBM P4)
Ram: 503 MB
Hard disk: /dev/hda (actually 150G SATA)
Michael Spencer Nova Scotia, Canada .~.
mspencer at tallships.ca /( )\
More information about the nSLUG