You may have noticed recently if you do a validation check on your sites code and it
comes back with four errors. This is Cloudflare’s injection of Google Analytics
on your site causing this issue.

I filled a bug report with them and once I get a reply
I will post here with either the final solution or a work around.

Apparently the issue has been fixed, cloudflare was “nice” enough
to not inform me and just deleted the bug report… *sighs*

Apparently while securing Inexistence (did an audit over the weekend) accounts,
I made an mistake of not adding false to the valid shells so pretty much all email was broken.

This issue has been addressed and all email accounts should work as expected.

If you were expecting an email and didn’t receive it I would email them and request them to resend it.
Thanks to George for the heads up about email being broken.

Sorry about this folks but hey you live and learn.

Once again we return to our normal scheduled program and enjoy!

So I did an svn pull a couple of days ago that apparently exposed some bad values in the config file.
I had a merry task of hunting it down, first I thought it was the imap server causing hiccups.
So I tried all sorts of crap including downgrading to an older copy of roundcube.

Finally I decided to diff -urN the conf files and just update everything.
And now I’m pleased to say cubemail works again for everybody.
(at least it should)

Sorry about the interruption of service but at least it didn’t take too long to fix imho.
Enjoy!

So tonight I went from the stable version of mysql 5.0.51a to 5.1.49 from debian backports.
*Sigh*

We’ve been having issues with database corruptions hopefully this will help out.
Please let me know if any of your sql powered sites have issues…

As always enjoy and I hope everything is great on your end…

Update – 10/18/2010
So the backport upgrade really didn’t help much…
I now have a custom shell script I wrote checking the databases every day around 4am.
If you site goes down around then don’t be surprised and a refresh of your browser should do the trick..
(It takes about 2 to 4 seconds to run and re-spawns mysql to fix the OOM issue it still causes)
I cannot wait till squeeze comes out (next version of Debian).

That is all…

All SSH access is currently denied for all users as there is a rather evil kernel bug:
Hole in Linux kernel provides root rights

Of course Debian hasn’t patched this in backports yet (which we currently use).
So until this is fixed please drop me a support ticket if you really need shell access.

Otherwise please be patient and I’ll let you know when I lift the ban…

Debian has fixed this with the latest backport kernel upgrade.
Enjoy your shell access again folks!

So I once again enabled QoS (I forgot to enable it when we came back to
using dd-wrt from the airport extreme).

For those who wonder what QoS is that explains it pretty well.

To explain it in easier terms I had to reserve part of the network so upstream
cannot be flooded causing a denial of service. Now mind you a very slight denial
of service more of a lag of service and this fixes the issue quite well. Sadly
this also causes another issue and that is people will now download slightly slower
from the server but everything will be more responsive so the trade off is entirely worth it.

The reason I re-enabled this service is because I’ve noticed several users sending very
large attachments (email) to yahoo. Now normally this isn’t much of an issue but yahoo uses
a deferral system the basically causes inexistence to keep retrying to send said email.
Which yahoo rejects until they decide when it’s ok for the email to be delivered.

(Yes we use DKIM and SPF so there is nothing I can do to accelerate it any further).

So for anyone sending email to yahoo and wondering why it takes so long for
them to get it, this is why and if you really care about the person you are sending
email to then I would strongly suggest to them to use a better service.

Gmail, Hotmail, MobileMe, and many more exist and do the job in a more
standards conforming manner that doesn’t piss off our server..

That’s all for now.

I did another apt-get upgrade mainly bind9 stuff nothing too important.

I installed imapproxy for squirrelmail/roundcubemail you should notice
an increase in speed when using them. (I know I did!)

Svn pulls were also performed tonight (roundcubemail, z-push)
Sadly because of so many changes I had to reset the roundcubemail
SQL database you will have to setup your account stuff again. (Sorry!)

I fixed mimedefang today for some reason it had a fit and prevented
sendmail from doing it’s job properly…

I updated all wordpress plugins on all of the sites I manage.

And I played around with WordPress 3.0 RC3 tonight it’s looking good.

Now friend connect with me via the bar at top and read the site more
often users! All major changes that occur on Inexistence I post here!

End of Line.

Today it was brought to my attention that inexistence.org’s domain name expired.
Less than 20 minutes later we are back up…

Sorry for the “slight” service interruption everyone.

Oh and if you still cannot connect to inexistence you really should be using opendns
by now. In other words go change your DNS settings so things work better!

The other day you may have noticed that you couldn’t view any websites hosted.

It was my fault I did an apt-get upgrade (which included php5 updates)
and the reason it broke is we use eAccelerator. And the issue was I didn’t recompile
and link it to the updated php5 of course I’ve done so since then.

So yeah did an apt-get upgrade
PhpMyAdmin is updated
svn updates were applied

And that is all for tonight…

I had to edit the sendmail access database today and remove a filter I put in there awhile ago:
From:hotmail.com IGNORE ZonelessDate

That entry created an issue that was blocking all hotmail emails to inexistence.
This has been rectified and I apologize to anyone that was expecting an email.

So in case you were wondering, yes hotmail users can email us again…