Subject: Hard time with IO bottleneck
From: Markus Fischer <markus@fischer.name>
Date: Mon, 23 Jun 2008 19:17:42 +0200

Hi,

I'm having a hard time with my system when it comes to IO performance.

As an example, I've three directories to delete with about 50000 files in 
each. I'm deleting on the "host" system already, using "ionice -c3" (only when 
idle) and CFQ scheduler is activated.

While this is done, I've about 11 vserver running.

As long as I have the delete job *not* running, the system has a load < 1.0. 
As soon as I start/resume the job, in short the load goes over e.g. 12 !

I'm currently completely puzzled as to what is going on here. Using vtop I 
can't find any two processes on top being the cause for the load. The list 
usually looks like

   top
   init [2]
   [kthreadd]
   [migration/0]
   [ksoftirqd/0]
   [migration/1]

continuing with the list of kernel threads and once in a while a process pops 
for a second to the top. I already ran vtop with a time interval of 0.l but 
the result was the same.

The delete job takes quite some time to finish, probably on the range of an hour.

Even if there would be a major IO bottleneck here with the discs (the whole 
host/guest system is located on a LVM on a Primergy RX200 S3) why would it hog 
the rest of the system down too, even with ionice idle scheduling?

I'm wondering if this has anything to do anyway with the vserver but I'm 
currently completely lost.

thanks,
- Markus