I just wanted to report that no trickles can be seen.

Have a nice day

---

Email sent.

I have the same problem with this task:
http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=17997390
The first two trickles appear good on the webpage. I found that the following had been sent by my BOINC client, but do not appear.

I confirm. The last trickles are from March 4.

Also, credits for completed tasks are granted according to trickles seen on tasks pages (according to trickles from March 4 and earlier only).
An example is here.

---

Have the same problem on two WUs. Here and here. WUs was completed like others, but from these two I didn�t get credit. I know, because of missing trickles, but where is the problem?

There was a previous batch of work that completed sending the zip files but not the trickle up messages. My understanding is that you will get the credits eventually but not until the programmers have time to fiddle with things. I notice that even for models producing trickle ups new credit isn't showing since before the weekend.

I suspect however that fixing this will have to wait until after the download server is sorted. If no one else has done it by the time I get back on my other computer I will email the project people.

Compare and contrast two tasks which I've finished recently.

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=17959489

Completed 23 February, and the task page shows 12 trickles received, the last one an hour before the task was reported complete. That's the normal way of things.

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=17992458

Completed 7 March, but the task page only shows 5 trickles - the last one on 4 March. I've been through my BOINC log - fortunately this was the only CPDN task running on the machine at the time - and it certainly sent the full 12 trickles. So it's the process which gathers up the trickle reports and puts them into the database (from where we can see them on the website) which has failed.

The first thing to do (after the download server is fixed) will be to get the trickle records up-to-date: only then can we determine if some recent v7.26 tasks are failing to send trickles, and convert the delayed ones into credits.

An ANZ model running on a Mac of mine is reporting contacts but not trickles, so I think we can assume it's a general problem - which should be resolved when the uploaded trickle files are ingested.

The trickle processing script had indeed crashed on 4 March. It's been restarted now, and the missing trickles are starting to appear on my sample task.

Note that the timestamp on the trickles as shown on the website is the time when they are processed, not the time they were sent by your computer, nor even the time they were received by the scheduler.

Exactly as has always been told here, the credits will be awarded as soon as the tasks are processed by the server.

Once again...

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=18145610

as an example.
Trickles lost or missing since March 23.

Have a nice day,

bonsai911

---

Hi,
Since 23 March no trickles on the page. Here is an example. All six models currently running report no trickles. Zip uploads are fine, but no trickles reported and few completed models after 23th have partial or no credit granted.

Yup no trickles reported -- probably viewed as a non-critical issue by the admins should they be aware of it (those minding the store on the message boards are no doubt aware of but are simply messengers).

No doubt the problem will get addressed with the priority that those who actually do the work choose to assign to it.

It's being followed up in the thread just below:
here.

Jonathan did try running the script yesterday. Will find out what went wrong on Monday probably.

Here we go again...

I'm running a "UK Met Office HadAM3P-HadRM3P Australia New Zealand v6.10" WU, and I just noticed that it seems them trickles are lost somewhere, since 29/30 April.

---

Professor Desty Nova
Researching Karma the Hard Way

Agreed, but this is a relatively benign problem: the models are sending trickles, and the trickles are being accepted by the scheduler. All that's missing is the script which processes the trickles and transfers them into the database for us to see. That can be restarted later, and will catch up (as it has before).

But... Monday is a public holiday in the UK, so it will be at least Tuesday, or possibly even later in the week, before anybody is likely to be in the lab to attend to it.

It seems to be that incident again...
cool down &keep up the good work,

bonsai911

---

Just to say it has been reported to the project people and an attempt has been made to restart the process but it didn't work. They will try again tomorrow.