Looks like I am still having issues with time zones. Everything I publish ends up as a scheduled entry.

(Attempting to fix that)...nope, doesn't work. The date of my local copy is 07:59:59 earlier than the server copy.

I don't quite get what it is doing.

I have been getting a relatively large number of spam bounces.

It has been quite some time since someone was sending mail forged from my address. Actually from close to my address. It seems that google doesn't count some variations ad different addresses.

e.g. bsmith and bs.mith at google are treated the same. I am getting bounces to the latter delivered to my mailbox which is in the format of the former.

I am getting annoyed by comment spam.

It had been a problem, then everything was working well again, now it appears that the spambots are able to recognize our challenge and do simple math.

I also know that the spam is not personal...but really. Spamming a post about my wife's brother being shot by a sniper in Iraq, so not cool.

Technorati Tags: spam

Adding another item of things not to do.

at now+30m
at> shutdown -g300 -i6 -y
at> CTRL-D

So I was testing some behavior to see if trying to allocate more swap than the available space would generate the message I was expecting (It does).

My next step was to use mkfile to eat up all but the smallest amount of space in /tmp that I could. Before doing that I set up the at job to try and recover if I ended up in a situation where all of my connections hung.

The result of the whole situation...

I was able to use almost all of my swap, the system was unresponsive for a minute or so while I was filling up /tmp.

While I continued to mess around seeing if normal tool use would also generate the message (It doesn't) I forgot to remove the at job.

So after making a note of my results and writing a message to that effect I remembered that I had an at job... unfortunately about 30 seconds too late to kill the running shutdown.

So much for my uptime.

Had a power outage at home this weekend. I wasn't there so I don't know how long it was.

I do however know that it was long enough to run down my APC 2200.

The problem here is that the old flaky hardware that I have been using for my main server appears to have lost it's powersupply.

sigh

lame

Everytime I post with BlogEd I get a java class exception because of an error making a cast. I didn't used to get that error.

I would post a screenshot but I closed it, I believe that BlogEd will facilitate adding an image after the fact.

So it was confirmed...sort of

Jun 01, 2005 01:11 PM 04:00:00

Jun 01, 2005 09:07 AM 04:00:00

May 30, 2005 01:09 PM 04:00:00

May 30, 2005 09:08 AM 04:00:00

My calls drop at 4 hours, still no idea where the limitation is and no real testing. If I can plan for another long call I will try moving the VOIP gateway outside the firewall but I can't currently think of a limit that would be imposed by the firewall that I wouldn't hit with other long running connections like music streaming. Maybe I need to try and start a Skype call and let it run for 4 hours and see if it drops.

So the new site go live... Current DNS lookups depending on who you ask return: www is a CNAME to CDN www is an A record to old site foo is a CNAME to www foo is a CNAME to CDN Which in addition to the above can return: foo is a CNAME for www is a CNAME to CDN is an A record for a CDN server foo is a CNAME for www is an A record to the old site This shows that the lookup for foo sometimes goes to a DNS server that has the old info for foo. Foo previously being a CNAME for www results in a lookup for www. The lookup for www sometimes goes to a server that has the new info and sometimes the old. The real confusion being the belief that if the final answer is the CDN there should be no intermediate CNAME. We just had a spirited conversation that resulted in browsers being restarted, which cleared up a number of interesting behavior patterns. Now to try and find some lunch since I didn't get a good breakfast and it is almost 3pm localtime.