Sussex Users Website Down

Message boards : News : Sussex Users Website Down
Message board moderation

To post messages, you must log in.

AuthorMessage
Michael DHEP
Project administrator
Project developer
Project scientist

Send message
Joined: 13 Jun 18
Posts: 156
Credit: 0
RAC: 0
Message 334 - Posted: 9 Aug 2018, 0:27:29 UTC
Last modified: 9 Aug 2018, 0:27:45 UTC

Hello,

The University users webserver (used by DHEP for serving class files) is currently down. (users.sussex.ac.uk)

We are looking into alternatives but hopefully all should be back to normal soon.

Kind regards,
Michael
ID: 334 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Gunnar Hjern

Send message
Joined: 9 Jul 18
Posts: 9
Credit: 32,440,374
RAC: 1,988
Message 335 - Posted: 9 Aug 2018, 11:17:07 UTC - in response to Message 334.  

Hi Michael!

I receive no Boinc credits, and the statistics pages all show that no additional work has been done.
Even the statstopology.php page show that no island topology exists!!
Nevertheless all my computers are working 100% on the tasks??

Will all this work (since the server went down) now be all in vain, or will we be acknowledged for it later,
when the server is back on line?

Kindest regards,
Gunnar
ID: 335 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
JugNut

Send message
Joined: 2 Jul 18
Posts: 24
Credit: 62,251,162
RAC: 162,942
Message 336 - Posted: 9 Aug 2018, 11:56:27 UTC - in response to Message 334.  
Last modified: 9 Aug 2018, 12:00:50 UTC

I recently had to shut down boinc to do some updates. But after resuming boinc over an hour and a half ago the DHEP WU's haven't started crunching yet. All the DHEP WU's on that box show 0% CPU usage & time so must not be doing anything.

I tried aborting three of them to see if the new WU's would run any better, but so far neither the new or old WU's on that box have started crunching and are all still showing 0% CPU usage. Oddly my other boxes seem to be running fine.

So is it the downed webserver that's causing the problems at my end? Or something else?

If it's going be a while I may as well put those cores somewhere useful for the time being? Any ETA?

All the best.
ID: 336 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Michael DHEP
Project administrator
Project developer
Project scientist

Send message
Joined: 13 Jun 18
Posts: 156
Credit: 0
RAC: 0
Message 338 - Posted: 9 Aug 2018, 13:02:09 UTC - in response to Message 335.  
Last modified: 9 Aug 2018, 18:09:27 UTC

Hi Michael!

I receive no Boinc credits, and the statistics pages all show that no additional work has been done.
Even the statstopology.php page show that no island topology exists!!
Nevertheless all my computers are working 100% on the tasks??

Will all this work (since the server went down) now be all in vain, or will we be acknowledged for it later,
when the server is back on line?

Kindest regards,
Gunnar

Hi Gunnar,

Thanks for your concern.

If the server remains up and the connection between the client and server is broken, even if it is for over a day, and the client reconnects: no work and no credit is lost.

If the server is down for all this time as was the case since last night: from the science point of view: no work is lost. However from credit tracking point of view you will earn less credit than if the server had been up. Apologies for this.

That said, even through alpha and beta stages and migrating from AWS onto a different Linux OS I hope you have noticedf there have been no server outages until this exteremely rare situation of a major institution's webserver going down!.. So even though we'd like to address this scenario, it is extremely unlikely to happen again and if it did, we'd rather focus on ways to keep the server up and running even if users.sussex.ac.uk was down rather than fix the 'credit allocation when server is down' scenario, which is nevertheless somewhere on our list of things to do. I wish we had more Java developers on our team! There are so so many cool funky things to code as well as critical server code efficiency and reinforcement improvements to get one's hands into!

Thanks for sticking around.

Kind regards,
Michael
ID: 338 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Gunnar Hjern

Send message
Joined: 9 Jul 18
Posts: 9
Credit: 32,440,374
RAC: 1,988
Message 339 - Posted: 9 Aug 2018, 14:58:36 UTC - in response to Message 338.  

Hi!

Thanks for a good answer!

Now, nearly all of my cores seems represented on the "statstopology", 45 out of 46 cores are now showing up there.

I guess we're back in business again, and that no further actions are needed?

//Gunnar
ID: 339 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Michael DHEP
Project administrator
Project developer
Project scientist

Send message
Joined: 13 Jun 18
Posts: 156
Credit: 0
RAC: 0
Message 341 - Posted: 9 Aug 2018, 17:08:30 UTC - in response to Message 339.  

Hi!

Thanks for a good answer!

Now, nearly all of my cores seems represented on the "statstopology", 45 out of 46 cores are now showing up there.

I guess we're back in business again, and that no further actions are needed?

//Gunnar

Indeed 👍🏻
ID: 341 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : News : Sussex Users Website Down


©2018