On my Munin graph for the "smartctl" plugin, it shows that since last month, the
Load Cycle Count has been continually dropping (
not sure if it was high in the first place...) on my harddrive.
It's dropped from about 128 down to 51.
Is a high or low load cycle count good or bad? This is on a
laptop running openSUSE 12.2 (x64) KDE.
I have had it sitting in Runlevel 3 (
for 8 days uptime) without a DE/WM loaded because it is acting as a pilot server for some services in the office and it's doing what it does very well.
It's primary services are vsftpd (FTP) and PostgreSQL/PostGIS (RDBMS).
I know a laptop is not designed to be left on for extended periods of time b/c one of the major problems is the Power Management features that a laptop uses for its hard drive. (
ie, placing it into idle mode when I/O's have been nil for some time, constant startups will wear the HD faster reducing life span/read-write performance). This power management can be turned off with
hdparm -B 255 /dev/sda. Currently it's at a laptop friendly setting of 128.
Code:
hdparm -B /dev/sda
/dev/sda:
APM_level = 128
smartctl -a /dev/sda
smartctl -H /dev/sda ...both report no errors
I don't have the exact details for the partitioning, but I know it's an encrypted LVM.
Code:
/Boot (sda1)
/LVM (sda2)
/(root) as ext4
/home as ext4
/var/lib/pgsql as ext4
/var/lib/pgsql/data/pg_xlog as ext3
NOTE: I'm using this laptop as a server for test purposes and a pilot project to convince the local government here to adopt open source solutions into their existing work flow. I've been promoting openSUSE. What's interesting is that we have a proprietary spatial database on an enterprise level server and it is not performing to par with my postgresql/postgis db on the laptop! Management is very impressed, given that our proprietary spatial servers were installed and configured by the companies that built the software and I set this up on my own...
No comments:
Post a Comment