Sunday, November 19, 2017

Quick weight loss update

So it's just a few days before Thanksgiving and I got on the scales to find I am now down to 318. Recall I started at 344 so that's 26 lbs to date.
I've been having to walk my mom's dog, as she's been out of town, so I have been hitting my step goal in Garmin Connect for the past seven days. 
Plus continuous calorie tracking in MyFitnessPal seems to be helping me make sure I keep my self on track to lose weight.
That's it for now, just a quick update.

Wednesday, November 08, 2017

freebsd postfix python policyd-spf ip addres validation error

The solution to my Problem was (for now) to add the following lines to the function _cidrmatch in /usr/local/bin/policyd-spf
def _cidrmatch(ip, netwrk)
  netwrk = unicode(netwrk)
  ip = unicode(ip)
  try:
    address = ipaddress.ip_address(ip)
This is done so that the python module ipaddr only has to deal with unicode string (as requested by the error message)
switching to python3 as suggested by other posts on the web did not work for me (maybe some missing dependencies?)

 Stolen from here, but it worked:
https://superuser.com/questions/1253657/freebsd-postfix-python-policyd-spf-ip-addres-validation-error/1253658

CIDR.pm

So for whatever reason the CIDR.pm file uses on my FreeBSD setup doesn't work for MailScanner.
II had to comment out these croak lines and now it works just fine
 
diff  /root/CIDR.pm /usr/local/lib/perl5/site_perl/Net/CIDR.pm
218c218
< The result is a two-element array: 
---
> The result is a two-element array:
437c437
<     
---
> 
596,597c596,597
<           #croak "Invalid netblock range: $r[$i]"
<               #unless $isipv6_1 && $isipv6_2;
---
>           croak "Invalid netblock range: $r[$i]"
>               unless $isipv6_1 && $isipv6_2;
794c794
<     my $mismatch = shift;
---
> 
801c801
<     if (!$isipv6_1 != !$isipv6_2)
---
>     if ($isipv6_1 || $isipv6_2)
803,805c803,805
<       return $mismatch if defined $mismatch;
<       #croak "Invalid netblock: $aa-$bb";
<    }
---
>       croak "Invalid netblock: $aa-$bb"
>           unless $isipv6_1 && $isipv6_2;
>     }
810c810
<     #croak "Different number of octets in IP addresses" unless $#a == $#b;
---
>     croak "Different number of octets in IP addresses" unless $#a == $#b;
979c979
<                                 
---
> 
1014,1015c1014,1015
<       next if _ipcmp($b[$i], $lo, -1) < 0;
<       next if _ipcmp($hi, $a[$i], -1) < 0;
---
>       next if _ipcmp($b[$i], $lo) < 0;
>       next if _ipcmp($hi, $a[$i]) < 0;
1284a1285,1286
>     $v =~ s/([0-9A-Fa-f]+)/_triml0($1)/ge;
> 
1291a1294,1299
> sub _triml0 {
>     my ($a) = @_;
> 
>     $a =~ s/^0+//g;
>     return $a
> }

Sunday, November 05, 2017

Gym Time

This morning I spent over an hour at the gym. 
I actually enjoyed it a lot, I noticed that I hard time not smiling.
I spent the entire on the treadmill and after 65 minute I covered 3.9 miles.
It was, according to myWellness cloud, my fastest pace for 3 and 5 kilometers, at 16:41 per mile.
At the end I was totally drenched in sweat, my shirt on the front and back, was totally soaked.
I also logged just over 8,000 steps on Garmin Connect.
I guess that's it really, but I just thought I would share.
See not everything I post is stupid work stuff, sometimes its also stupid me stuff.

One Massive Screw Up

So at work we have some MX servers running Postfix with Mailscanner/SpamAssassin on FreeBSD.
Last week I decided to continue my FreeBSD 10.3 upgrades and decided to tackle all five of these machines at the same time to 11.1.
Everything went well, or so I thought.
Turns out somewhere in the upgrade process when I upgraded all the ports the MailScanner port decided to over-write all the config files and so we would up with a very malfunctioning mail scanner.
Plus due to some other strangeness, the servers were eating up all of their memory.
Well I wasn't sure why, but nailed it down to MailScanner.
So I reinstalled MailScanner and all of its dependencies.
That didn't do anything.
It was next figured out that a configuration file had been renamed, but the original still existed so a quick copy fixed that error.
Then an issue was discovered in the antivirus scanner used by MailScanner, this was resolved by making a quick change to a configuration file.
It was then discovered, that the Bayes database had grown to large to be useful, in excess of 5GB and it was trying to load that into memory.
The database was purged and an automatic process was put into place to purge the database once a week.
Finally, as an extra precaution the main MailScanner configuration file was modified to reduce the maximum number of children processes running at a time.
 This limits the amount of mail that can be processed, but also ensures the memory used didn't exceed what was available
Anyway, all of this kept the mail for being delivered for a few hours.
I don't have anything technical to report about it, just this horrible synopsis.
The bad part is I still have a bunch of servers I need to upgrade.

Oh in coincidental and slightly related news the migration from self-hosted Exchange 2010 to Office 365 is ongoing, but working smoothly.  I guess I shouldn't be surprised.  The documentation is rather exhaustive, both official and otherwise.

I managed to migrated my own mailbox as a "proof of concept" and before I move any others I'll be testing it like crazy.  Of course because I am not changing the mail flow until we're all off Exchange 2010 I need to make sure that the MXs continue working as expected.