pitter patter on the keyboard

Something about this new “apps” paradigm ClearOS has entered has been gnawing away at my subconcience ever since I wrote my little critique on 6.3. In a post on the ClearOS forum regarding missing IPsec support I think I was able to finally articulate the off-ish smell that has been driving me mental.

kfox:

I can’t seem to find the IPsec app for ClearOS 6.3

I see the paid “dynamic vpn” app in the market place and it appears to reference an independent IPsec app.

The Dynamic VPN app is an extension to ClearOS’s IPSec VPN app. The service allows IPSec to be used in situations where either one or both of the gateways are on a dynamic IP address issued by the ISP or in cases where instability using unmanaged IPSec tunnels exists.

Herballizard:

http://www.clearfoundation.com/docs/release_info/clearos_community_6.2.0/final_release_information

PITA

kfox:

The unmanaged IPsec tool has been unmaintained for a few years and was dropped in version 6. It’s open source, so if someone wants to revive unmanaged IPsec, go right ahead.

Yeah I love the whole “It’s OSS, you do it if you like it so much!” attitude at the same time architectural decisions seem to have become increasingly marketing-driven. If it was too much trouble to update the old IPsec module why not cut out all the paid bits of the for-profit Dynamic VPN app? Smells a little fishy.

Maybe I will make it. If you hire me. Unfortunately, I have to put food on the table and the people who pay for my time have very little use for a webconfig interface once I have it rolling. Being someone who has contributed little more than some help on the forums and a couple VM images I wouldn’t be so whiny if this wasn’t a functionality ClearOS didn’t already have at one point.

I’m beginning to question the logic of continuing to use ClearOS when I have to do so many things myself; I’m a Gentoo admin so it goes without saying that I love to do everything myself – but I use this crazy, neat little redhat system because it used to save me countless hours and let me respond to network crises quickly.

It feels like the foundation has cut off its nose to sell its face. A lot of stuff seems to be missing or half baked just so they could roll out this new “Marketplace” paradigm in time for RHEL 6. A paradigm which itself rubs me all sorts of wrong ways.

It’s a shame they gambled on buzzword dollars rather than building on an already great platform. I hope I’m dead wrong; that the gamble pays off and we end up seeing a whole bunch of quality third party “apps” from the community but the sad truth is that functionality was always there and we didn’t see a whole lot of participation back in the day (and I’m not pretending to have been any help!).

On the surface, it looks like this new app framework was designed mostly with the intent to make it easier for paid services to be integrated. I wonder which kind of apps the Foundation staff members will be focusing most of their attention on now. They certainly don’t seem worried about the lack of a free IPsec app despite every crappy embedded router’s support for it and highly critical Advanced Bandwidth rules have been bumped two versions (so far!).

Oh well, I know only too well that we all gotta make that dolla. Maybe the corporate makeover (and hopefully increased revenue that follows) is what Clear needs to propel itself to new heights of greatness. I sincerely hope so.

UPDATE You should really read the thread; Dave Loper did a great job of explaining why things have gone this way and what the path forward looks like. I’m a lot more optimistic now.

I’m sure anyone who works for ClearFoundation and sees this will think “well, you can’t please everyone!” but this is a list of things they have managed to horribly screw up that every admin needs to know before plunging head first into ClearOS 6.3 or worse, an earlier 6.x:

• You can no longer limit or reserve bandwidth for a whole IP or IP range. I know the documentation says you can. You can’t. This was apparently done in 6.2 then carried into 6.3 to make bandwidth management play along with multiwan – something that seemed to be possible for years until now. From the app’s review tab:

  by Asad Siddiqui – June 20, 2012

  Following modifications are required;The bandwidth limitations are on network card interface only, there is no option of limiting bandwith on the basis of single IP address or range of IP addresses. Although this was provided in Clark Connect.This option may kindly be added in this application

• Speaking of apps: WTF, apps? 6.x is clearly the Foundation’s idea of caching in on “cloud” and “app” clichés. Software is no longer packaged, it is apped and you don’t get your apps from a repo you get them from the marketplace. It’s like they’ve tried to make a routing distribution appeal to a twelve year old girl. Routers are not something that should be built by people who have no concept of routing and making it approachable to those who do not is aggravating to those who do. This is not a desktop distribution, why are they trying to broaden their target demographic?

  I see what you did there. I couldn’t disapprove more.

• There is no way (at least that I’ve found so far) to uninstall apps in the webconfig. You must feel around and guess the package name then yum erase it.
• Every god-damned page in the webconfig now has a huge, unnecessary app column taking up valuable space in the hopes that they might up-sell you on the commercial apps you already passed up on install. Fsck off, I’m trying to configure my firewall – I don’t give a crap how many stars it has.
• Even more of the setup process is done in the web interface now and god help you if you happen to put in a wrong name server or you may find yourself wondering why it’s taking forever to not time out when it looks for new packages.
• Registering with ClearSDN is now mandatory; you’re SOL if you think you can set up the router without an active connection and drop it in later. Wonderful. ET phone home!
• Kernel-devel doesn’t actually contain the kernel sources and kernel-sourcecode is missing. You have to do it the hard way:

  wget http://mirror2-houston.clearsdn.com/clearos/community/6.3.0/dev/SRPMS/kernel-2.6.32-279.2.1.v6.src.rpm
  rpm2cpio kernel-2.6.32-279.2.1.v6.src.rpm > kernel.cpio
  cpio -idmv < kernel.cpio
  cd rpmbuild/SOURCES/
  cp linux-2.6.32-279.2.1.el6.tar.bz2 /usr/src/
  cd /usr/src/
  tar xjf linux-2.6.32-279.2.1.el6.tar.bz2

• “Development Tools” package group has been replaced with clearos-centric clearos-devel. This pulls in 170 packages meant to help you design apps and whatnot but mostly useless if all you need is a C build environment.
• No more free IPsec! There is still a paid-for “Dynamic VPN” app which provides this functionality but the old IPsec module has been dropped for good.

Doubtless I will have plenty more nasty things to say about this new major version as time goes on and it reveals its sins to me. Do check in from time to time.

The only nice thing I have to say is way to go on the 64 bit version – now if only they would provide Xen images in addition to every other virtualization platform that should NOT be used to run a router…

RENDER UNTO BETA WHAT IS BETA’S.

Wall Street Journal contributor Clint Boulton gaffed in a blog entry Monday titled “CIOs Beware: New Macbook Pro Will Be a Bandwidth Hog.”

Clint argues that larger screens take up more bandwidth, apparently forgetting that there is a difference between screen resolution and the actual resolution content is delivered in:

…Better quality displays require more network bandwidth, which allows users to increase data consumption. Consider that experts told CIO Journalearlier this year that the new iPad, which includes a Retina display of 2048-by-1536 resolution with 3.1 million pixels, would slow enterprise networks to a crawl and increase data costs from carriers. Now imagine how a Macbook with 5.1 million pixels — two million more than the new iPad — will increase data traffic in office networks.

CIOs would do well to monitor network usage and make sure their employees aren’t watching too much high-definition content on YouTube and other data-hungry websites. CIOs whose policies for content consumption are lax must be prepared to increase bandwidth. Another option might be for CIOs to require workers who want to bring their own high-powered devices to the office to bring their own bandwidth as well. At the very least, CIOs might want to follow the lead of companies such as Google, which give employees a monthly “bill” for the IT services that they consume, and make the usage a matter of record throughout the company.

Apparently Clint is doing everything he can to meet his publishing quota, this bollocks is a continuation of his March 22 article, “The New iPad Could Create High-Speed Headaches for CIOs”

The rotten “experts” (with an s) in all this seem to be Amtel CEO P.J. Gupta who “sells software that sets alerts and notifications on bandwidth consumption.”

One wonders how this technically challenged sap managed to get a gig writing articles for Chief Information Officers when he can’t tell the difference between a sales pitch and objective analysis. I can see the HR people at WSJ are top notch.

Despite all the corrective comments the article hasn't been pulled or edited.