That's some qwerty slaping bombasticly good news -a-roo!

WCG website is down for me, but WUs are still flowing.

They are getting better at communication... The site now says

Webserver is currently down due to an issue with RabbitMQ in the Openstack hosting environment, hosting is working to recover the webserver now.

BOINC services are still up, so far only the webserver was affected causing the load balancer to return this 503 error page when volunteers attempt to access the WCG Website or Forum.

In reply to unixchick's message of 30 Apr 2026:

They are getting better at communication... The site now says

Webserver is currently down due to an issue with RabbitMQ in the Openstack hosting environment, hosting is working to recover the webserver now.

BOINC services are still up, so far only the webserver was affected causing the load balancer to return this 503 error page when volunteers attempt to access the WCG Website or Forum.

Well, I doubt that "hosting is working to recover the webserver now". Sharcnet hosting only works normal office hours, and it's now 18:45 in Toronto.

It's back. That was unusual, since Sharcnet so far has not worked outside of office hours.

A new update too:

April 30, 2026
Website outage approximately 17:30-23:00 UTC April 30th, 2026 was caused by an issue with RabbitMQ in our Openstack environment. The webserver became unresponsive, and failed to recover after reboot. Hosting intervened to recover the instance, and we were able to bring the Website and Forum back up - as of 7:15pm EDT.

In reply to Grumpy Swede's message of 1 May 2026:

That was unusual, since Sharcnet so far has not worked outside of office hours.

That's odd, given their purported role from what I read at wikipedia. Heck the last two organizations I worked for provided systems support during normal office hours, 24x7 in special cases, and even in normal times there was always someone carrying a beeper/phone to provide off hours/weekend support.

In reply to bill's message of 1 May 2026:

In reply to Grumpy Swede's message of 1 May 2026:

That was unusual, since Sharcnet so far has not worked outside of office hours.

That's odd, given their purported role from what I read at wikipedia. Heck the last two organizations I worked for provided systems support during normal office hours, 24x7 in special cases, and even in normal times there was always someone carrying a beeper/phone to provide off hours/weekend support.

And i expect those organisations paid well for that support.
The usual response to "How much support do you want?" is generally "How much can you afford?"

---

Grant
Darwin NT.

Can't disagree with you. But since both related to (US) national defense, uptime is/was paramount. The wikipedia article for SHARCNET says, in part, "It was formed to allow members access to larger, faster, and more modern computer resources than they would otherwise be able to afford, and to retain researchers at their organizations", Now if I were a member of that consortium I'd want support for that resource during off hours also. Since the uptime seems to run less than 24x365 and, at best, "office hours" of support is less than 24% of the time.

From: How long should I expect to wait for support?

How long should I expect to wait for support?
Unfortunately Alliance/SHARCNET does not have adequate funding to provide support 24 hours a day, 7 days a week. User support and system monitoring is limited to regular business hours: there is no official support on weekends or holidays, or outside 9:00 - 17:00 EST .

Please note that this includes monitoring of our systems and operations, so typically when there are problems overnight or on weekends/holidays system notices will not be posted until the next business day.

And now the Jurisica Operational Status page, at https://www.cs.toronto.edu/~juris/jlab/wcg.html, is down with the following information:

Due to a scheduled University power outage, the resource you are looking for is temporarily unavailable. The power should be restored early Monday morning, May 4th. Please return to the main departmental page by following the 'Home' button above or clicking this return link.

We apologize for the inconvenience.

On a positive note...it allows them to do an orderly shutdown, for a change. FOR not OR...Darned autocorrect.

In reply to bill's message of 1 May 2026:

Can't disagree with you. But since both related to (US) national defense, uptime is/was paramount. The wikipedia article for SHARCNET says, in part, "It was formed to allow members access to larger, faster, and more modern computer resources than they would otherwise be able to afford, and to retain researchers at their organizations", Now if I were a member of that consortium I'd want support for that resource during off hours also. Since the uptime seems to run less than 24x365 and, at best, "office hours" of support is less than 24% of the time.

Quite, just cost $50 Million and it can’t be fully supported all the time. Just hope those 134,000 processors and several hundred GPU’s power up Ok.?

In reply to Hadrian's message of 2 May 2026:

Quite, just cost $50 Million and it can’t be fully supported all the time. Just hope those 134,000 processors and several hundred GPU’s power up Ok.?

Datacentres are like boats and aircraft- as much as it costs to buy one, the costs of running them are so much worse...

---

Grant
Darwin NT.

Just an observation .... WCG has now posted Stat's on 7 separate days. While not seven days in a row ... Anything is wonderful after that 180+ day gap.

I guess the Wish's and Hope's finally drifted down into the Electronic gears.

Bill F

---

In October 1969 I took an oath to support and defend the Constitution of the United States against all enemies, foreign and domestic;
There was no expiration date.

What does this mean? Wed 06 May 2026 08:47:58 AM CDT | World Community Grid | Tasks won't finish in time: BOINC runs 94.7% of the time; computation is enabled 100.0% of that.

You set your cache too high. BOINC Manager does a benchmark (you can start manually when you think it is wrong) and depending on the estimated computing time of workunits it looks if it gets done in deathlinetime when boinc runs this 94,xx%

---

----------Boincing since 2011-----------
DSKAG Austria: http://www.dskag.at

BOINC sometimes gets things very wrong in the time it estimates tasks will take to finish. In the past, micromanaging several projects has been particularly good/bad at messing up the predicted times. These days, I mostly only run one or less often two projects at once on my machine and they still overestimate a lot on one of them.

When you attach to a project, BOINC relies on it's Benchmark results and the estimate provided by the project for how much work is required to process a Task.

If the benchmarks haven't been run, then it relies on the default values.
The project supplied estimates for the amount of work required to process a Task can very between extremely good, to beyond ridiculously bad. Ideally, the project should overestimate the amount of work required so that an extremely slow system won't initially end up getting more work than it can handle before the deadline is reached.

Once at least 10 Tasks have been processed and Validated, the Manager will then use those processing times to work out the Remaining (estimated) times each time it downloads new work (although this has been broken for a while now with the current Manager releases).

But even with all that; while BOINC keeps track of how long it takes to process a Task, how much Time BOINC is running when the system is up, how much processing it can do when BOINC is running, your Resource share settings between projects and your cache settings; the biggest issue is systems that are being heavily used by other software (eg video encoding, Folding@hokme, playing with local AI models etc). BOINC has absolutely no awareness of this, and it's only by the big difference between Runtime and CPU time for CPU work that it can be observed, and it takes a while for BOINC to adjust to the much longer Runtimes than what they should be before it can bring things under control.
And the larger your cache, and the more projects you do, the longer it will take for things to settle down.
(0.5 days and 0.1 additional days is plenty. If you have a reliable project then 0.02 days and 0.1 additional days is plenty).

---

Grant
Darwin NT.

Once at least 10 Tasks have been processed and Validated, the Manager will then use those processing times to work out the Remaining (estimated) times each time it downloads new work

And, if you have a project that keeps bringing out new versions of task types before you get to ten, then it never changes from the initial estimate!

Be nice to know when we stop “flying blind” and can actually track and see our results etc.