Fixed: Some users couldn't post this morning

[drmike]

So,  this morning some folks were unable to post.  The group of people should have been those who posted last overnight inside a time window.  No user group, just based on activity of those users --- all of them.

IPB (forum software) was giving odd long times before they could post again --- an administrator set time between post sort of message.

After consulting with MannDude, determined the origin of the problem was a bad server set time.   Manually poked the right time into place and all seems to be back to normal.  Giant_Crab was impacted and able to post now.

IPB = grrrr.   Their support = grrr slow.

To do: ntp synching.

 

[MannDude]

Thanks for noting the timezones were all goofed.

Last night (this morning) I was testing a failover solution so I killed the VM for 10 minutes... (failover solution didnt work... its on my to-do list) so when I brought her back to life and saw everything come online I went to bed for 2 hours. I woke up and noted several PMs from people saying they couldn't post and that they had to wait six hours.

I did not limit anyones permissions despite the error. The server came back with a botched timezone and it goofed things up. After it got fixed, it goofed up some results in the 'View New Content' but I believe it should be more or less resolved now.

Report errors here.

 

[mikho]

I noticed that I got 2 Pm's from The Mann with the some content.


And that there were "no new content" for a long time today.

 

[drmike]

Yeah @mikho, Mann butterfingered his PM yesterday and a subset of the users got a double dose.    First one had a typo/spelling thing...

Content - new lacking was the time issue.

 

[MannDude]

I noticed that I got 2 Pm's from The Mann with the some content.

I sent out a mass PM to all the members, but I accidentally stopped it when it had only sent <250 out of 1200+... With no real way to start it where it left off, I had to restart the process. So some of you got two, everyone else got 1 and got it with the correct spelling of ".org" instead of ".og" that was used in the first one sent... haha.

Don't worry, I think this is the first site-wide one I've sent and only did it to talk about 2013. Hopefully won't send another one out like that for a long time!

 

[AuroraZero]

Yeah suuuuuuuuure it was a glitch   I bet MannDude was cold from running outside naked or something.

 

[MannDude]

Yeah suuuuuuuuure it was a glitch   I bet MannDude was cold from running outside naked or something.

IPB Staff got back to me to confirm that was the cause.

If the server reboot caused the server time to be less than the latest recorded timestamp for a number of functions, that can cause some very odd behavior.

So, keep that in mind. Do not roll the time backwards, either purposely or by mistake on an IPB install.

 

[mikho]

I sent out a mass PM to all the members, but I accidentally stopped it when it had only sent <250 out of 1200+... With no real way to start it where it left off, I had to restart the process. So some of you got two, everyone else got 1 and got it with the correct spelling of ".org" instead of ".og" that was used in the first one sent... haha.


Don't worry, I think this is the first site-wide one I've sent and only did it to talk about 2013. Hopefully won't send another one out like that for a long time!

Only read the second one, I thought that was the TL;DR version.

 

[drmike]

Note to all: keep your server time/date synched.

I've been down this road before.   IPB broke only slightly.   Certainly are other software packages that break worse.

 

[Shados]

Do not roll the time backwards, 

Yeah, this pretty much always results in bad things on computer systems. Also on causality.

 

[Magiobiwan]

Freaking time paradoxes. I hate it when those happen.

 

[Ruchirablog]

1 step closer to the cluster

 

[fisle]

Now fix the Git, all repositories are empty

 

[MannDude]

Now fix the Git, all repositories are empty

It's being upgraded. Should be back soon.