I finally got the new app working ok
Bu since the outage last weekend have huge problems connecting to the server.
With that i mean forum, site and the unit distribution now i don't know why since other seem to have less problems with it.
I cleared my dns caches but that did not help and still see hundreds of messages like these in boinc:
15-9-2009 18:11:29||Project communication failed: attempting access to reference site
15-9-2009 18:11:29|Milkyway@home|Temporarily failed upload of de_constrainted_15_3s_6_855509_1253025761_0_0: connect() failed
15-9-2009 18:11:29|Milkyway@home|Backing off 1 min 0 sec on upload of de_constrainted_15_3s_6_855509_1253025761_0_0
15-9-2009 18:11:29|Milkyway@home|Temporarily failed upload of gs_constrainted_82_3s_6_895301_1253028670_0_0: connect() failed
15-9-2009 18:11:29|Milkyway@home|Backing off 1 min 0 sec on upload of gs_constrainted_82_3s_6_895301_1253028670_0_0
15-9-2009 18:11:29|Milkyway@home|Started upload of de_constrainted_82_3s_6_909142_1253029699_0_0
15-9-2009 18:11:30||Internet access OK - project servers may be temporarily down.

---

Its new, its relative fast... my new bicycle

Im getting the same connect errors maybee Travis is working on the server but the whole site has been slooow today. Havent had any problems up till now

---

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

---

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

It just got a lot better some minutes ago!

Hope it stays that way.

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

It just got a lot better some minutes ago!

Hope it stays that way.

I reduced the feeder sleep time to 1 second (down from 2). Sadly I think this is the most we can make the feeder work, so if we go through any more extensive optimizations or have a large influx of users we're really going to need to get another server machine.

---

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

It just got a lot better some minutes ago!

Hope it stays that way.

I reduced the feeder sleep time to 1 second (down from 2). Sadly I think this is the most we can make the feeder work, so if we go through any more extensive optimizations or have a large influx of users we're really going to need to get another server machine.

How about calling it Milkyway_gpu? :p

---

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

It just got a lot better some minutes ago!

Hope it stays that way.

I reduced the feeder sleep time to 1 second (down from 2). Sadly I think this is the most we can make the feeder work, so if we go through any more extensive optimizations or have a large influx of users we're really going to need to get another server machine.

Could I suggest that consideration be given to increase the task count? The reason I am contacting the server every 20 minutes is because I only get 23 tasks per hit. I know we are on a tightrope here ...

Or back to the concept of longer tasks or packed sets though that is going to obviously mean changes to the code again ...

At the least I have only one system hitting on ya'll ...

Thanx this gave me enough to access the site again so i guess you where right about the extra. And yes i think the server gets stressed by the constant hannering of the hungry machines ( ofcourse mine not ;) ) outthere who wants more and more units xD
Now that i can only get half the points i did before i try to get more units so i maybe can get my old rac back.
So CP your new app did not help much i have now half the rac, sadly the guy i was competing with just bought 2 new 4890 cards and thats for me simply impossible. So becoming one of the top guys in boinc is out of my reach again

---

Its new, its relative fast... my new bicycle

Get the extra server anyway. You'll need it sometime, so better to get it, get it set up and running smoothly to it can take over or load share with the old one under normal circumstances rather than have to do it all under pressure when the old one collapses.

---

Cheers,

PeterV

.

I think the new version of the GPU app is making the server pretty strained again. I'll see what I can do but I'm not sure how much faster we can get her going :D

It just got a lot better some minutes ago!

Hope it stays that way.

I reduced the feeder sleep time to 1 second (down from 2). Sadly I think this is the most we can make the feeder work, so if we go through any more extensive optimizations or have a large influx of users we're really going to need to get another server machine.

So can't you start asking for donations again for new equipment? Did the last rally cover the expenses until grants are received? Never heard if the project could stay running until the grants were received.

Maybe the new server could be on Milkyway GPU?