AliGrant.com
 
Valid XHTML 1.0!

Valid CSS!

Powered by PHP

Get FireFox

1&1 Internet


Archived News

Opensuse 13.2 logrotate timerThursday 26th March 2015
I was slightly surprised to notice that my /var/log directory has bloomed in recent months. Clearly logs weren't being rotated. logrotate is a program that will archive, compress and prune old log entries. It's a must if you have logging enabled on your system to avoid unlimited growth.

Traditionally a cron-job was used to run logrotate daily, but it seems in Opensuse 13.2 that this has vanished. Instead it has been replaced by a Systemd timer, which works just as well, but you may not have it enabled by default. In order to do this run:
systemctl enable logrotate.timer
systemctl start logrotate.timer

You may also want to kick off a manual rotation instead of waiting for a day to tick over:
logrotate /etc/logrotate.conf


MP3 vs CD qualityThursday 12th March 2015
There is a never-ending debate in the pseudo-scientific world of perceptible audio quality. The argument is that MP3 is a lossy compression format and so it doesn't sound as good as the original CD.

The lossless part of the MP3 encoding works around the psychoacoustics of music - i.e. what our psyche can't perceive doesn't need to be retained. We know there is a upper and lower limit to what we can physically hear, but then there is a lot more clever goings on about sounds close together that we won't be able to perceive and such. The up-shot of all this is that we can take a large CD and compress it down to a much smaller size without it sounding any different to us.

The argument though, is really by how much can we compress it. Some people play safe and go with 320kbps, which is MP3's maximum bit-rate; but I feel this is sort of missing the point of MP3 and you may as well go with FLAC at that point. A few years ago everything seemed to come in 128kbps CBR format, but it seems the current de-facto is 192kbps VBR. The Variable Bit Rate instead allowing more bandwidth to be used in complex parts and to save it in the less complex parts, providing an average across the length of the track.

There are numerous articles out there on the ever-so reliable Internet about what bit rate should be used and there are some tests using various methods such as A/B and double-blinds - all of which show that there is no point in really going above 160kbps. Yet there are people who will swear that it's all rubbish - mostly in the crazed audiophile world where people will pay a small fortune on cables because somebody suggested the electrons will be happier.

Whilst I'm the first to jump on any bandwagon against outlandish claims, I'd like to point out something that has been bugging me in regards to these tests. Some of them are crowd-sourced, some aren't. But generally in all the cases we have no idea about the listening environment in question - and this is something that audiophiles do know: there is no point in any of it if you don't have a good enough playback device.

How do you listen to your music? Off a mobile phone? PC speakers? A dedicated amp and speaker set? Stand-mounts? Floor standers? Where's your cross-overs? Is your EQ balanced? How "bright" is your amp anyway? Who did the acoustic treatment in your room, and how much reverb is correct?

The questions get progressively more silly (although not to some audiophiles). When I first discovered MP3 it hadn't acquired it's catchy name, it was just MPEG 2 Layer 3, and I used a Fraunhofer CODEC to compress an uncompressed WAV down to a much, much, smaller WAV. It was so good I could put clips and such onto floppy disks! Yes. 1.44MB floppy disks. And with no loss in quality! Well... of course there was a loss in quality, in fact, there was a massive loss in quality. To fit stuff down to that size the sample rate was halved and the bit rate was dropped to 56kbps. But - on my everyday home PC I could not tell the difference in CD quality and MP3 because the speakers I used were so very, very rubbish.

When I finally got better speakers I noticed a significant difference, and settled on 128kbps. As I have got older and slowly upgraded my equipment I've noticed I've had to increase the quality of MP3s in order to bridge the gap to CD, even though my hearing will have been slowly getting worse at the same time. My last upgrade to my home hi-fi made me think that there may well be something in the argument from audiophiles, the problem is, the rest of us are on such poor equipment that we can't tell the difference.

I would like to see a good study done of self-confessed audiophiles on high-end equipment in high-end listening rooms to see if there is a possibility that the differences can be identified.

192kbps seems to work fine for me at the moment, but I know that my lounge is full of harsh edges that cause echos and that certain frequencies pool. I know I can improve the sound of my lounge without changing my speakers. My equipment is very much entry-level, and a bit of a hodge-podge. I wonder if I had really good kit if that quality gap would become apparent again?


BizTalk Delay shape not delayingThursday 5th March 2015
I had an odd issue earlier where the Delay shape in a BizTalk Orchestration was not delaying. The timeout happened seemingly immediately (100ms) - when I checked the code it was hard-coded to a two minute timespan.

The code is well established so nobody had been changing it recently. I figured it was a funny five minutes and restarted the host instances but that didn't help.

As it was time related I had a hunch to check the time, and I noticed that one of my hosts' time was drifting. It turned out a DC had had its time configuration messed up. The issue remained once the app hosts had synced up which lead me to check the SQL database host, which was also suffering from drift.

I resynced the SQL host (w32tm /resync) and the problem cleared itself up.

Keeping time synced across your systems is very important, the code here started to fail but before long Kerberos would have started to fail preventing logon events.

In a Windows domain, all clients will sync with the DC (or, one of) unless you explicitly override it. Be sure that your DC is syncing with a time-server.


Previous Next