I have seen such high failure rates in the wild - cross batch, cross manufacturer even.
But invariably, they were proceeded by a thermal event - I have never, ever seen worse than 10% failure in a datacenter that has a clean aircon record, and would expect 5% or better unless there were power issues too.
if you are seeing that sort of failure rate, I would be giving special care and attention to any "service visits" the ups or aircon guys may have made in the two months prior to the problem starting.

MariaDB is not much if any better - Ok, I can see his original point - he shared the source to MySQL so that he could get the benefits of community bugfixing, but retained the commercial rights so that he could sell commercial usage licences and still make money.

I can also see how, when offered a buttload of money by SUN, he could get up front and in one lump sum what he might make in years of normal trading - and SUN, having no db solution of its own to compete, was as good a new owner as any.

However, with MariaDB he is trying to have his cake and eat it too - he wishes to start a new "community" edition of MySQL so he can still steer the project, despite having taken his pieces of silver and ran once already. Despite (or even because of) his "experience" in running the MySQL project, I would not consider him a particularly good choice to control a fork.

Not sure if this is applicable, but - many torrent trackers accept donations via paypal....

Actually, that plays to a second danger. If you can get a court order like this, then presumably at some point they can convince a Belgian judge that "Official Belgian newspapers" should automagically get a higher rating on news.google.be than foreign/unofficial ones... Google search results could end up ordered by lawsuit rank not pagerank :(

Google is a private, foreign, totally unaccountable organisation.

Clearly it is a private, foreign, but {within the jurisdiction of Belgian courts and accountable to said courts} organization, or this article wouldn't exist.

A Belgian judge has said "Remove all content from all your sites, but in particular, for google.com and google.be, for articles, images or graphic representations of the newspapers bringing this case"

Search engines work by indexing the content, comparing the index with the search terms, and using that to generate results. No content = no search = no result in the list. how is that hard to follow unless you are a Belgian landshark looking for cash damages not actual results (given robots.txt, as is repeatedly pointed out, can let you fine tune what google does or doesn't show)?

Why not look into contributing to an Open Source solution in your chosen area? see if you can find (for example) some PM tools that almost are good enough, and make them better?

This is largely an issue with the "front end" - dynamic changes to iptables don't auto-write themselves, but that is true also for (for example) Cisco IOS. it used to be that you couldn't even insert a rule in an ios access list (you had to append, or failing that, blank out the whole list and start over) but like IPTables, you can now insert and delete from the list on-the-fly.

A competent front end should write "hot" to the loaded list, but also update a static file so that they can be re-loaded on reboot. iptables has a built in "save" method that can generate such a file, but you don't always want to commit every change to the startup config.- but blaming the engine for the poor quality of coding involved in what is only a pretty front end onto a very competent packet filter is a bit unfair.

Interesting. Walmart took over a UK chain called ASDA some years ago, who has a scheme for doing this - staff go and "pick" the goods from the shelves in a real store, just like a normal shopper would, bag them up, do a CNP transaction for the payment, then ship them out in a van to the homes.

Maybe some stuff does flow upstream?

Cisco's support is very good - its expensive, but you get what you pay for. I am not sure why that means their kit is noticeably any better though; you tend to find, particularly with their higher end kit, your choices are a) pick which set of bugs you can live with or b) go with a beta that is near untested In fairness, (a) is usually good enough; MS have set the bar for "enterprise level" so low nobody expects perfection. But still, its not cheap, and *having* to have bought from an approved vendor, and have bought support, in order to get bugfix patches is a major pain. What is really needed is for someone else to up their game to match the level of support.

I would say that the desktop should be the *last* thing you replace.

First come the servers - migrate from windows or unix to linux, keeping the front-end as similar to the original as possible. Main problem here is if you are a exchange and outlook site, most exchange replacements are simply not up to scratch, or cost as much as just going with exchange - you want an outlook front end, with support for folder sharing and calendar busy searches, and in most cases that just isn't going to happen.

Next come the apps - and in most cases, this is easier. As you upgrade and replace, select those which have cross platform clients that look pretty much the same on both linux and windows; if they are in fact entirely platform independent (java?) so much the better, and if you can arrange for them to be web deliverable than better still.

Finally you can swap out the desktop - on the given day, the start button becomes a penguin button - but gives pretty much the same menu, with the same apps, looking and acting the same.

However, all this may change - the new target is no longer the server or desktop, but cloud computing - at which point, it no longer matters what you have front end or back, provided the front can render and the back run whatever cloud format comes out on top.

Seconded. The purpose of this is not to hinder your use of the files, but to make them traceable back to the original purchaser if they are shared online. as you don't have a licence to share online, that's fine - nobody but you should be in a position to care what data is in those files.