Hello, I am running SuSE 9.0 Enterprise, latest kernel
2.6.5-7.147-smp, on a IBM Blade server. It has 2 Xeon processors, 2,5
gigs of RAM, and it is attached to a FastT store.
I have create a 1,2 terabyte raid0 partition (on top of fastt's raid5)
and handle *several* image files (tif, soon to be djvu) with sizes
around 70k. most of the files are on a single directory, and all is
working fine (uptime is 89 days), but since I converted from reiserfs
to XFS I am getting crappy backup thruput, but I am thinking that this
is CA's software fault, since it has been patched.
Anyway, I have a 1gig swap partition, and it got filled up quickly,
while I only run plain Apache2 (worker) to serve the images (no PHP or
other dynamic content). and there's nothing that seem to be using this
swap area, but it is 100% anyway, I after while created a swapfile,
also 1gig, and it also became 100% full, usage growth rate is about
400 megs/hour.
RAM is ok, see below:
any ideias ? Just started happening with XFS...
Thanks in advance.
alvaro@blade01:~> cat /proc/meminfo
MemTotal: 2594792 kB
MemFree: 73116 kB
Buffers: 40520 kB
Cached: 2347360 kB
SwapCached: 220 kB
Active: 1031532 kB
Inactive: 1384416 kB
HighTotal: 1703860 kB
HighFree: 3712 kB
LowTotal: 890932 kB
LowFree: 69404 kB
SwapTotal: 2097120 kB
SwapFree: 0 kB
Dirty: 1776 kB
Writeback: 0 kB
Mapped: 32948 kB
Slab: 79368 kB
Committed_AS: 2851264 kB
PageTables: 672 kB
VmallocTotal: 114680 kB
VmallocUsed: 21712 kB
VmallocChunk: 76392 kB
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 4096 kB
|