Server outrages

[gd_barnes]

Fixed. Equipment issue. Sorry about the problem.

[AMDave]

Complete backup set offsite download completed for 2014-05-02. (4.8GB)

[AMDave]

Complete backup set offsite download completed for 2014-05-16. (4.8GB)

[AMDave]

Complete backup set offsite download completed for 2014-06-13. (4.8GB)

[mdettweiler]

PRPnet port 9000 will be going down shortly, for up to an hour, while I upgrade it to PRPnet v4.3.6.

See here for more details. This is a simple upgrade that we've done many times before and it'll only be putting this port on par with all the others. (Gary approved this upgrade more than a year ago, maybe two years, I just haven't done it yet. That'll be fixed now... )

[mdettweiler]

...and it's all done. As expected, no problems so far. Everything is fully backed-up, so if we run into any problems we can roll back. (I don't expect any problems, since we've been running this version without issue for upwards of 2.5 years on multiple other servers.)

@Gary: FYI, the backup files are located at:

• Database backup: /home/max/prpnet9000-db-backup-20140619-2046.sql.lz (use "plzip -d [filename]" to decompress; use "sudo su max" and enter your password to log in as me if it says you don't have permission to read the file)
• prpnet9000 directory backup: the folder "(DELETE AFTER 1 WEEK IF NO PROBLEMS)prpnet9000-backup-20140619" on your desktop - name is self-explanatory

[mdettweiler]

All noprimeleftbehind.net services appear to be down presently - I can't reach the web site, any of the PRPnet servers, or SSH.

The most likely explanations are:

• Server IP change that needs to be updated in DNS (though I think Dave might have that set up automatically now)
• Power outage at Gary's house, in which case we probably won't hear from him until it's back up

[AMDave]

NPLB unplanned outage confirmed.
Either a power or comms outage at the server location.
Commenced shortly after Sat, 28 Jun 2014 19:11:59 -0500.

[Lennart]

Originally Posted by AMDave

NPLB unplanned outage confirmed.
Either a power or comms outage at the server location.
Commenced shortly after Sat, 28 Jun 2014 19:11:59 -0500.

It started with the servers I could not reach them but I could reach all other webpages.

Lennart

[mdettweiler]

Originally Posted by Lennart

It started with the servers I could not reach them but I could reach all other webpages.

Lennart

Maybe the web pages were still cached in your browser?

Given the time of year (thunderstorm season in the U.S.), my bet is on it being a power outage...but, I can only speculate until we hear from Gary. We've had all sorts of crazy stuff happen to the servers in the past (I can scarcely believe the number of motherboard blowouts Gary's had - both on his AMDs and Intels, which were bought at different times, and even many of the replacement boards*).

*To be fair, the failure of the Intel boards can be attributed directly to my own bad advice; when I recommended the components to Gary for those builds, I had neglected to check the motherboard manufacturer's website to make sure it could handle the right wattage. Turns out, that model was only designed for Core 2 Duos and lower-wattage Core 2 Quads, of which the Q6600's we were using were not one. I'd put together one of my own builds from the same parts at the same time, and it blew out in the exact same fashion. Moral of the story: always check the manufacturer's website, the label "Core 2 Duo/Quad" on Newegg is not enough.

As for the AMD boards, I have no idea why we had such bad luck on them, except that both Gary and I have now learned well to avoid buying the very cheapest motherboards that meet the requirements.

[mdettweiler]

BTW @Gary: when you turn the PRPnet servers back on, you might want to set the reservation time limit to 1 week temporarily. IIRC, PRPnet refuses results returned after the deadline (even if they could otherwise be accepted) - this wasn't too big a deal in the past but now that we're doing much larger tests (particularly at CRUS) we'll want to watch this.