Orion: anyone else having post-crash issues?

[Gwaihir]

As the announcement says, Orion was failing last night ("night" by my European clock). When I looked back in this morning it supposedly was back up for hours, however my account still wasn't. That seems (nock on wood) to have been fixed right after I reported it this morning, but it still leaves me with some questions.


1) What happened? Clearly it wasn't purely a hardware replacement. Perhaps along the way someone tried to fix something that turned out not broken after all? I can easily understand it if such happened; I know first hand that trouble shooting can be a messy business. However, I do want to know what happened, what sofware was touched, so I can do some specific double checking to make sure all is indeed fine now, rather than wait what may or may not turn up over the next few months (and always at an akward time, thanks to Murphy).

I got dissapointingly stonewalled when I asked support this. I have just been told for the third time that nothing has been done to any of the software / data / settings. This time that was even stressed as obvious. Yeah, right, as if my account stays down all by itself |:(. So, I've come to wonder out here now, both still looking for the answer and to vent my annoyance a bit.

If you have found anything wrong with your Orion-based account after last night, please post it here to inform me and others on the machine on what was affected. In my case my unique IP wasn't connected to my account anymore.


2) Howcome the techs considered it finished without noticing my account was still down? I even had a ticket in. Isn't such verified, rather than just presumed to work again? (And it wasn't exactly hard to verify: nothing worked: no web, e-mail, SSH, simply nothing, as my domain was pointing into an unconnected void.) I find it rather dissapointing that due to such carelessness a two hour problem was turned into a 12 hour one for me.

Taking it one thing at a time, I haven't asked support this question yet, but I do hope folk will look into it and make sure it doesn't easily happen again.

[JPC-Masood]

1) You have got the correct answer as no software changes were done.

2) Please pm your ticket # and I can take a look at it.

[Gwaihir]

Hi, thanks for dropping in. That's ticket #12423822 (you already took a peek).

Yet somehow my IP, and probably others, as the tech spoke of "IPs", were no longer connected to their accounts on the server.. Oh wait, so that wasn't the server not responding to them, but the network not getting the packets to / from the server anymore? If so, I finally get the "nothing software-wise was touched" , but I'm not less puzzled. How did that happen, as well as go unnoticed? Isn't the network stuff supposed to be so reliable that it's 100% uptime, guaranteed ?

Also, shouldn't me having a ticket in have helped? I.e. when a tech answers a ticket like that with "yeah, probably this known issue that we're working on", isn't it set aside to be verified once that issue is resolved? If not, then consider that a recommendation for improving procedures. You really need to test yourselves; whether it's because of a busy work day or a good night of sleep, it can easily be hours before a client has a chance to do.

[JPC-Masood]

The issue was related to unique IPs only, which were not responding. Apparently they did not bind due to some bug in cpanel. We have already looked into that but can not verify until the server is rebooted next time.

[Ron]

Masood,

I've had a similar "dedicated IP only issue" once before (a LONG time ago). Might not a "ping" of some sort be set up on a dedicated IP on each machine?

Just pondering.

[Gwaihir]

Yeah, I posted a suggestion like that last night in response to Jag talking about improving the monitor script(s).

[Ron]

They can try to load my forum's index page as often as they like. Once every 5 minutes would be 288 times a day, which wouldn't be noticed.

It would check connectivity, webserver, PHP/CGI and MySQL. Not bad.