[PLUG] more information

Tim tim-pdxlug at sentinelchicken.org
Thu Mar 29 03:41:30 UTC 2012


> { This is very odd.  Can you try pausing the cp process temporarily (^Z)
> to see if your system becomes more responsive? }
> 
> I did so, and I was able to regain control and normal execution rate of my
> commands almost instantly.

Hmm, not even with 'fg' ?

> The drive is a SATA drive which plugs into the motherboard via a SATA
> backplane.  This is an expensive Supermicro 7046GT-TRF motherboard.

And you were copying to and from locations on the same drive?  Or was
there another drive involved?

What filesystem(s) are mounted at these (source and destination)
locations?  (ext3? ext4? other?)


> { There's a possibility that cp or a library it depends on has a memory
> leak, but I would find this unusual.}
> 
> Unfortunately after killing the job and then killing the shell script that
> was running the cp commands, "top" shows
> 
> top - 20:04:44 up  2:04,  4 users,  load average: 0.33, 1.75, 3.05
> Mem:  24733672k total, 24447148k used,   286524k free,   314712k buffers
> Swap: 25165812k total,     1040k used, 25164772k free, 22884668k cached
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> --------------------------------------------------------------------
> 
> So there is no load on the system, but almost all the memory got taken away
> and it is not being returned as one would expect.

If cp and the associated shell are truly dead, then it sounds like
this could be a kernel bug.  Perhaps in a filesystem driver.  With
more information, we may be able to help you track down known bugs
associated with these.

tim





More information about the PLUG mailing list