...might be worth a thread to get more people to use the "Daily Schedules" feature to limit communications on Tuesday if they aren't in the likely to run out of WUs during a maintenance shutdown group.

Ok, I will byte :)

How do you use the local manager settings for the network over-ride, to shut down communications? I am getting the impression that the "time" needs to be based off of UTC rather than the local time equivalent?

Since my local UTC is -5 or -6 and they go down about 8am and come back up about 3pm (or earlier)[on short days] by my local time, what should my "day/time" setting be?

Tom

You use your local time zone in the configuration (use the 24hr format)

I am at UTC -5 too and this are my settings: (in the global_prefs_override.xml file)

<day_prefs>
      <day_of_week>2</day_of_week>
      <net_start_hour>22.00</net_start_hour>
      <net_end_hour>8.00</net_end_hour>
</day_prefs>

It stops the network around the outage start time 8 AM and return after 14 hrs at 22:00 hrs (10 PM).

But i run with the spoofed builds so my large cache handle the long delay to restore the network access.

In your case, to return at 3PM as you posted, your must use: <net_start_hour>15.00</net_start_hour>

I was using the "GUI". And I had it start a 07:00 and end at: 21:00.

It was 22:10 hours when I checked and the uploading had been suspended. So I am wondering if the GUI is broken?

I have inserted the code with the 15:00 hours. But it hasn't shown up on the GUI. I am going to restart BOINC Manager and see if it shows.
It showed up.
It is strange. The first column is the Network startup time and the second column is the Network shutdown time.
But it would have made more sense to me, if the first column was when the network shutdown and the second when it starts up again. So I tried that, and it didn't work.
Talk about "counter intuitive". I think the columns should be labeled, at least.

Tom

---

A proud member of the OFA (Old Farts Association).

Definitely still an issue with uploads
[edit] Who'd have thought posting a message would help :) [/edit]

. . That's just the SETI ghost playing wid ya!

Stephen

:)

I thought after preparing the brewroom (for another long session tomorrow) and having a feed at the top pub (along with a few beers of course) that the upload problem would've gone away,

but I was sadly wrong. :-)

Cheers.

. . Welcome to the seedy world of server caused grief like the rest of us mortals ...

. . <sigh>

Stephen

:(

Einstein still does. Shouldn't be bothering Seti. I have <max_file_xfers>8</max_file_xfers>
<max_file_xfers_per_project>8</max_file_xfers_per_project>
which should be enough to handle my 3 projects.

I was curious and tried it out on my fastest box since it was still very backed up currently. It worked ok for about 20 minutes and then started doing re-try's and "pending" again. I have set it back down to 2 for now but it is in the cc_config.xml file now.

Tom

---

A proud member of the OFA (Old Farts Association).

Yep, rather difficult to get new work when you can't return existing work.
Something for them to work on tomorrow.

---

Grant
Darwin NT

It has finally caught up for me (small amount of work). The validated WUs are now in the proper section, and not stuck in the "in progress" section.

Uploads still sticking, but enough getting through (or at least not timing out) at the right time to get replacement work every now & then.

---

Grant
Darwin NT

Definitely still an issue with uploads
[edit] Who'd have thought posting a message would help :) [/edit]

. . That's just the SETI ghost playing wid ya!

Stephen

:)

Just when I thought I was out, they pull me back in!

---

While often a bit slow, uploads appear to be making it through without the need to retry now.

---

Grant
Darwin NT

While often a bit slow, uploads appear to be making it through without the need to retry now.

. . For me they seemed to pick up a bit about 7:30am UTC.

Stephen

:)

I was using the "GUI". And I had it start a 07:00 and end at: 21:00.

It was 22:10 hours when I checked and the uploading had been suspended. So I am wondering if the GUI is broken?

I have inserted the code with the 15:00 hours. But it hasn't shown up on the GUI. I am going to restart BOINC Manager and see if it shows.
It showed up.
It is strange. The first column is the Network startup time and the second column is the Network shutdown time.
But it would have made more sense to me, if the first column was when the network shutdown and the second when it starts up again. So I tried that, and it didn't work.
Talk about . I think the columns should be labeled, at least.

@Tom

If you enter the changes directly into the file you need to select to read the local prefs files or your changes will not work until you restart the client. Nothing is broken with the GUI you just need to tell the client to read the changes.

Yes the GUI is "counter intuitive" making a little mess in our head when you try to input the time. In the first column you put the starting time (21:00 in my example) and in the second the stop time (7:00) exactly the inverse of the outage times. Remember you tell the client when is NOT to use the network, not the inverse. So they work as expected if you look from this side.

---

all stats on the user and host stats page seems to be about 3hrs behind

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

all stats on the user and host stats page seems to be about 3hrs behind

Or 10,800 seconds.

The SSP says that the replica database is 10,262 seconds behind the master. These two facts might be related.

Ah nice catch.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

How do you use the local manager settings for the network over-ride, to shut down communications? I am getting the impression that the "time" needs to be based off of UTC rather than the local time equivalent?

Since my local UTC is -5 or -6 and they go down about 8am and come back up about 3pm (or earlier)[on short days] by my local time, what should my "day/time" setting be?

Tom

Rather than mess around with scheduling I just use Boinc Manager itself. Activity tab then click on suspend network activity. I re-activate 12-14 hours later.

Seems a much simpler way to deal with things

The only difference is if you do manually you must babysitting the host, i prefer to make the things more automated.
That's why i use the function already build-ed on the Boinc itself.

---

How is it babysitting? When it gets dark, you switch the lights on, in daylight you switch them off. Job done.

How is it babysitting? When it gets dark, you switch the lights on, in daylight you switch them off. Job done.

If you have a single host close to you that could work, but if you have more or in a different places, not an easy task even with remote access. Now imagine those who have small farms. They certainly like to automate all the tasks.

In my case, since my host is at my home and i spend most of the week on the farm with almost no internet access, make things automated is mandatory.

---

MB splitter output has fallen off quite a bit, but that could just be a result of the huge amount of AP work that's been going out for a while now.

---

Grant
Darwin NT

Well, if that fresh batch of GBT doesn't hold us for the weekend, I don't know what will.
But the recent AP feast has indeed been pleasant! Even enticed me to add a couple more GPUs to existing crunchers.

---