gpg: failed to create temporary file '/var/lib/lurker/.#lk0x576b1100.hantslug.org.uk.17453': Permission denied
gpg: keyblock resource '/var/lib/lurker/pubring.gpg': Permission denied
gpg: Signature made Fri Jan 19 11:06:20 2007 GMT
gpg: using DSA key 1E38DD6257A4363C
gpg: Can't check signature: No public key
On Fri, 2007-01-19 at 10:53 +0000, Jim Kissel wrote:
>
> Rob Malpass wrote:
> > Hi all
> >
> > To what extent is a PC as fast now as it was when first bought?
>
> Wow! What an apples and oranges question.
>
You're making me hungry!
> >
> > As we know, hard disks get fragmented over time. I'm honestly not sure
> > how much this affects Linux as much as Windows or any other OS, but the
> > general thinking is that if you defragment a slow Windows filesystem,
> > performance will improve. I've personally never seen it improve
> > Windows that much - but it does have some effect.
>
> Although there are some disk defragmenters for Linux/Unix, they never
> have enjoyed much popularity. Probably due to the fact that most modern
> Linux/Unix file systems are good as housekeeping and keep fragmentation
> to a minimum.
>
As I understood it, Linux based filesystems are not as susceptible to
fragmentation as FAT/FAT32/NTFS due to the different cluster selection
algorithms used. IIRC so long as a (for example) ext3 filesystem has
plenty of free space it shouldn't get into a badly fragmented state.
> Having said that, if your application is very disk I/O intensive, a
> new/clean disk may improve performance.
>
Indeed faster spindles, multiple spindles, spreading load to reduce I/O
contention, all these are sensible performance tuning measures unrelated
to disk fragmentation.
> The periodically re-install of Windows use to be due to DLL "rot". A
> re-install cleared up this problem, but my experience with this is
> ancient history on Win95. I can't comment on the more "modern" versions
> of MS-OS as I don't use them.
>
It still happens. I have seen plenty of XP system run "like a one-legged
dog swimming in treacle" which miraculously speed up after a reinstall.
I guess there's a few reasons for that including the user not actually
re-installing all the crud they previously had, data being less
fragmented due to new apps being installed and nothing having (yet) been
deleted and so on.
> I new HDD, could be faster. Better and more cache, higher rotational
> rates, faster seek times. If you replace the HDD with exactly the same
> spec drive, I wouldn't expect to see any change in performance.
>
Unless you re-install as mentioned above. If you used something like dd
to copy the contents bit-for-bit from the old disk to the new one I
would expect performance to be the same because the layout (and hence
fragmentation) would still be there.
Cheers,
Al.
