The Slamlander

Glenn Fry … Smuggler’s Blues

Slamlander posts should not be hidden behind a link anymore.

GIGABYTE – Support&Download – Motherboard – CPU Support List – GA-6VX7-4X.

Motherboard support link.

On 15Dec09 my Websites all went offline. They are all Wordpress or MediaWiki sites and I was getting Cannot find RDBMS errors from them. Note that I religiously have the Web Applications on a different host than the RDBMS¹ .

Mine is a complex system. I have an internal DNS TLD and about 14 external SLD² . I also use NET10 addressing³ behind a NAT wall. It is a fairly impenetrable system.

On 15Dec09 I did a maintenance update of the DynDNS client. DynDNS created a new service called Internet Guide for small users on a simple LAN without their own DNS resolvers. Unfortunately, they configured their DNS Updater client to have it on by default. My maintenance update installed it and it reconfigured my NIC to use the DynDNS resolvers. I run this thing on my Web Application server. Of course, this broke the hell out of everything because the DynDNS resolvers had no clue where my internal RDBMS host is.

I had no idea that it had done this. All I knew is that it was broken I started bug shooting. This morning I finally, in desperation, started looking at NIC settings and discovered the name server address changes from the DynDNS client. I looked at the addresses and traced them back to Internetguide.dyndns.org.

Anyway, a long story later and my websites are back online. This article is part of the test.

Life meme

I have a server that’s been doing yeoman’s work since I got here. It’s a 200MHz Pentium in an old AT chassis with 110GB of HDD but only 256MB of RAM. The memory isn’t upgradable. It was one of the original two machines that SG had when I moved here. Out of the six towers we have, it is the slowest and least functional.

While it used to hold our main file repositories, those repositories have been moved elsewhere. This makes Condor completely redundant. While I thought that I could put slakware or another Linux variant on it, the limited RAM capacity puts too large a crimp on such an installation. In the light of the 900MHz ATX MoBo that I have in storage, it would not be cost-effective to upgrade or retain Condor.

I will strip Condor of parts and dump the chassis and MoBo. I will then obtain an ATX chassis and a DVD drive, combine those with the ATX MoBo and Condor’s parts and build a new Franken-box.

I will not install Win2KAS on it though. The new OS will be Linux. It will need to live within my Active Directory networking environment and I understand the SUSE Linux Enterprise Server is supposed to be able to do that. Alternatively, I am also looking at OpenSuSe Linux and Samba.

So as you all can see, I am about to dive back into the depths of tinkerware after an absence of almost seven years. The more I look, the more I am afraid that nothing much has changed there since I left it.

While I appreciate the benefits of a lower than expected employment report from the US Federal Government, I detect a serious discrepancy between the labor department report and the ADP report.

While the Labor Department¹ reported a drop in non-farm job losses of just 11,000 the ADP report of the day before was 169,000. That is a discrepancy of 158,000² or 1436%. ADP has never been that far off the mark before.

Granted, the ADP report is usually a bit different from the Labor Department data. However, this is because ADP does not have quite all of the data. Nevertheless, ADP has never been this far off. ADP collects the data from its client database and as they are the premier payroll outsourcing firm, their data is usually within the few standard deviations from the Labor Department report. The fact that they are off by 1436% is more than normal in may indicate an error on the Labor Department Report.

There is a story here, I think. In the past twelve months, ADP has gone through an extreme effort to make their report more accurate. That they are now, for November only, this far at variance raises some concerns with me. Yes, I was also caught up in the exuberance of the good news. However, I woke up this morning troubled and began investigating³.

US Bureau of Labor Statistics (USBLS), US Department of Labor (USDOL)

The US Bureau of Labor Statistics publishes the official Employment Situation report. It collects most of its data from Unemployment Insurance claims and is therefore considered more accurate. However, this also has its caveats;

1. Those who have exhausted their employment benefits but are still not employed are simply dropped from the database. They are still unemployed but are no longer counted as such⁴ .
2. UI claims are only one of the tools, which are backed up by survey data, which is also statistical in nature and subject to the same statistical errors as the ADP Report.
3. The methodology is much more opaque than that of the ADP Report.

ADP

ADP (Automatic Data Processing)  is a private firm that is an outsourcer for payroll deliveries and other HR services.

Automatic Data Processing, Inc. (Nasdaq: ADP), with nearly $9 billion in revenues and about 570,000 clients, is one of the world’s largest providers of business outsourcing solutions. Leveraging 60 years of experience, ADP offers the widest range of HR, payroll, tax and benefits administration solutions from a single source. ADP’s easy-to-use solutions for employers provide superior value to companies of all types and sizes. ADP is also a leading provider of integrated computing solutions to auto, truck, motorcycle, marine and recreational vehicle dealers throughout the world.

Among their services, they publish a National Employment Report. The details of their methodology are published online as is the report itself.

The essential thing to remember is that this data is gleaned from the payroll data of ADP clients. The caveats are;

1. ADP does not have 100% market share and therefore their data suffers from statistical errors when extended to the entire jobs sector.
2. People leave companies for many reasons and not all of them result in filings for Unemployment Insurance.

ADP publishes their report one day before the official US BLS report.

Analysis

It is good that we have two different reports trying to derive the same data. However, given the nature of the source data, the ADP reports should be slightly more optimistic than the USBLS report. When it is not then both reports should be compared and contrasted to derive why they differ⁵ .

This month, the USBLS report and the ADP report differ wildly, as I stated above. My surmise is that one of them is broken.

It shows the fickleness of the markets; On one US job report ….

1.018 Swiss Francs (CHF) per US Dollar (USD/$)

1.511 Swiss Francs (CHF) per Euro (EUR/€)

1.676 Swiss Francs (CHF) per British Pound (GBP/£)

November Jobs Report is positive at only 11,000 jobs lost v 125,000, with important revisions for October. This just drove the US Buck up to

0.9908 Swiss Francs (CHF) per US Dollar (USD/$) (Sub-Parity)

1.5082 Swiss Francs (CHF) per Euro (EUR/€)

1.6712 Swiss Francs (CHF) per British Pound (GBP/£)

Regarding the voter results for the minaret ban in Switzerland. 57% of the voters in 22 cantons voted in favor of the ban.