Twitt

QA testing has finally successfully finished and all the bugs have been resolved. Production environment is being tested right now.

Considering the unexpected issues we ran into, we prefer to test it a couple of days more.

We will provide further details in the next few days.

Few days...

Well...

We were unable to bring our production environment to the same state as the QA environment this week. As we also have yet to resolve an issue that prevents BOINC clients from downloading workunits, the effort to bring the Grid back online has stretched into next week.

Seems a basic thing.

I wonder if they understand the difference between the meta-descriptions (workunits, tasks) and the required files? The client doesn't "download" a workunit: that's something that we probably understand better than they do. But without a functioning message board, there's no way to help them out.

Well, I'd say that the sentence could well be shortened to "I wonder if they understand"

Anyhow, for those with deep knowledge about the inner workings of BOINC, getting in touch with Dr. Igor Jurisica, who is
responsible for the Krembil takeover of WCG, could be pretty easy. Perhaps offer som help, something that IBM obviously
haven't done. I'd say they are stuck, and it's a lot they do not understand.

Contact page for Jurisica: https://www.cs.toronto.edu/~juris/jlab/contact.html

Edit: Or at least tell them to start the forum, so they can get help in that way.

Krembel should contact the Official Google boinc_dev mailing list/forum (if they already have not done so).
It's For people developing, debugging or porting the BOINC software (client, server, and Web).

for those with deep knowledge about the inner workings of BOINC,

If you're looking at me, I'll have to sit this one out. I'm a long way from home (touring Portugal), and will spend three of the next five days on a succession of trains to get home - up to 12 hours a day. And when I get home, I'm going to sleep....

WCG has always been a bit of an outlier in the BOINC world, so the raw BOINC documentation won't help much (and it hasn't been very well maintained for recent changes in the last few years) - this will have to be approached at a conceptual level, rather than RTFM. I've been invited to a Google meet for BOINC developers about an hour after my last train lands, so I might drop in on that and plant the idea.

If WCG are still stuck, and the trains run on time, that is...

Thanks. If I was in a position to chat with them reasonably quickly - which I might be on Wednesday and Thursday, when I'm spending long continuous periods of time on well equipped trains with power and wifi - I think I'd suggest that one of the researchers took a machine home (or borrowed their wife's/husband's): tried to join WCG or their QA project; set the http_debug event log flag; and took the resulting log back to the lab the next day for analysis. It's more likely a connection error, not a transfer error per se.

I've sent a copy of the lack-of-progress note to the convenor of the BOINC meeting, and suggested that it might be helpful for someone experienced to act as a "conceptual mentor" - someone to help them find the right chapter and page in the manual, before they start reading the nuts and bolts.

This is just a wild a$$ guess but... Maybe this is why IBM abandoned WCG. They had more important things to do:

IBM-powered Mayflower robo-ship once again tries to cross Atlantic

Well, it's Tuesday May 31 and there's still a few hours left in the calendar day. In North America.

Krembel still hasn't posted any news on Twitter, Facebook or project website about the fixing of WCG.

Seriously makes me wonder how well they are going to actively handle a living, breathing project of this magnitude?

Working through the undocumented foibles of a new cloud migration can be an 'adventurous' project in itself.

(Don't believe the cloudy Marketing hype...)


Keep searchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Ouch!...


Just had two of my physical servers physically yanked over to a different datacentre elsewhere in London...

No idea if the entire rack was forklifted or if the individual servers were carried individually...

And the aftermath is...

By chance, one of the servers was already shut down for an upgrade so all is benign for that one.

However, the second server was still live and picked up the signals for a graceful shutdown... Except when powered up again, the startup failed safe to not expose the services... There was an uncomfortable long pause to activate the new network routing and then a few hours to discover the one missing feature after the move... Now all running as before with less than a millisecond difference in the connections.

All part of the fun of relying on someone else's infrastructure...


To be fair, that has been the first time in over two years. But then... First times always involve lots of coffee...

Enjoy!
Martin

(Yes... I'd missed the maintenance email warning about the move... Immediately afterwards noticed and realization dawned that one moment too late...)

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

What is another missed deadline since it's now over 3 months since the "migration" started?

Half the day is gone in North America. It looks like Krembil will go without another update today and this week.

New update on the WCG Website., and it sure doesn't seem as if they are even close to restart the project.
Nothing on Facebook, or Twitter yet though.

WCG June 3rd Update

I'm probably an old cinic, but the way I read it is:

We are having technical problems so we have decided to expand our marketing department.

They probably have to spin a lot of bad news in the upcoming period.
Poor intern ...

... They probably have to spin a lot of bad news in the upcoming period.
Poor intern ...

So, this is delayed start #7. Will lady-luck be with them or will it just be another crap shot?

My bets and odds fall on the latter, rather than the first one.

Yep. 9th Nov only to crash a week later on the 18th Birthday challenge.

This offloading by IBM seems similar to WCG.
IBM CEO explains why he offloaded Watson Health: Not enough domain expertise

Well it's the start of a new week, will the gerbils at Krembil actually restart WCG even partially?

My bets are on more technical issues being found and little information being presented to us about what broke this time.

I suspect IBM dropped gave away the White Elephant project probably because it wasn't making any money.

Krembil took on the project and is now deep in the muck trying to fix what sounds like proprietary software without free IBM support.

I suspect IBM dropped gave away the White Elephant project probably because it wasn't making any money.

Probably true, but that was by design - it was under IBM's philanthropic wing.

Still not a sign of life from WCG, or any update since June 3rd.

The sound cricket chirping is deafening...
As of today, 2022-06-14 @ 1735 EDT (2235 UTC) no updates since 22-06-03. The gerbils must have fallen off the wheel and can't get back up

Well, it's Thursday evening 9 pm EDT (0100 UTC Friday) and the cricket chirping is deafening to the point we still don't have anything from the Krembil gerbils since June 3.

Wondering if the problems are so severe they don't have time to even advise the "New Public Relations" critter what to say spin.

It's now mid-day Tuesday, June 21 Eastern time and the WCG gerbils must still be on summer vacation in hibernation...

It has been over 4 months since WCG suspended sending new tasks and almost 4 months since the "shutdown" started.