I'd love to know the trick to clear the stalled uploads. Nothing I have tried has allowed them to upload. Several thousand uploads pending on each host.

Not sure about "thousands", but I just use the "transfers" tab in Boinc Tasks and a lot of mouse clicks!!

Saying that, all my Linux machines have cleared their backlog, however both Windows machines with less uploads take forever.

Saying that all my Linux machines have cleared their backlog, however both Windows machines with less uploads take forever.

Since the servers also run Linux, is that an indication that - deliberately or not - 'Linux shall speak unto Linux', and foreign dialects are harder to understand? :P

This is TORTURE... SETI style...

+1

Stephen

Fun and games on the left coast I see. Upload server having a very difficult time handling the huge volume of uploads from the noise bombs. Out of cpu work again, gpu cache continues to drop because stuck uploads prevent downloading of any new work.

. . Completely out of work here :(

Stephen

<sigh>

Fun and games on the left coast I see. Upload server having a very difficult time handling the huge volume of uploads from the noise bombs. Out of cpu work again, gpu cache continues to drop because stuck uploads prevent downloading of any new work.

. . Completely out of work here :(

Stephen

<sigh>

Hmm. Almost full caches across 6 machines here

Saying that all my Linux machines have cleared their backlog, however both Windows machines with less uploads take forever.

Since the servers also run Linux, is that an indication that - deliberately or not - 'Linux shall speak unto Linux', and foreign dialects are harder to understand? :P

Could it be something to do with TCP 1323, for timestamps and window scaling?

Linux has all of that enabled by default, and I know I remember before the servers moved to the co-lo, windows machines had tons of problems with the saturated link up the hill, but once timestamps and window scaling were enabled in the registry, comms got much better.

https://docs.microsoft.com/en-us/previous-versions/windows/it-pro/windows-2000-server/cc938205(v=technet.10)

Also found an old thread from this forum when searching for 1323..
https://setiathome.berkeley.edu/forum_thread.php?id=71002

I had *some* issues with uploads earlier today, but they did clear up and go through, but it is also a linux machine, so.. instead of having a failed connection after a few seconds, the connection would stay open for several minutes, with data transfer going to 100% and taking 60+ seconds for the server to send an ACK.

timestamps and window scaling certainly help, but they're not a complete and total fix

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Fun and games on the left coast I see. Upload server having a very difficult time handling the huge volume of uploads from the noise bombs. Out of cpu work again, gpu cache continues to drop because stuck uploads prevent downloading of any new work.

. . Completely out of work here :(

Stephen

<sigh>

Hmm. Almost full caches across 6 machines here

. . My Linux boxes are getting some work now but still having problems with uploads on and off. My Windows boxes are completely stalled and OOW.

Stephen

:(

Yes, I remember RFC 1323 too - somebody (not me) made a good find there, and I was happy to confirm it - it was exactly what we needed at the time.

But that need was different: it was about downloads, when packets were dropped on an established, but congested, link.

But this problem is about establishing the link in the first place - getting a connection. And there are very few packets to drop in an upload. There may be (almost certainly is) an equivalent RFC defining the connection handshake, but I don't know how to find it offhand.

When a beer doesn't work, have a sleep instead. ;-)

Cheers.

Getting downloads, but the upload rate isn't fast enough. The problem for me seems to be so many of these noise bombs that I'm working through. They finish in 3 seconds and take a lot longer than that to upload.

---

I'd love to know the trick to clear the stalled uploads. Nothing I have tried has allowed them to upload. Several thousand uploads pending on each host.

Not sure about "thousands", but I just use the "transfers" tab in Boinc Tasks and a lot of mouse clicks!!

Saying that, all my Linux machines have cleared their backlog, however both Windows machines with less uploads take forever.

Clicking Update does nothing since each machine has a constant 8 tasks Active with all the rest pending. When a task does eventually upload, it sits at 100% and then waits from a minute to a half hour before getting an ACK.

I run spoofed clients, so "thousands" is accurate and not some exaggeration. The Threadripper host has 3300 tasks in Upload currently.

[Edit] Finally out of work on the Threadripper host. Off to do work for Einstein and MilkyWay and wait for all the slow Seti uploads to clear so I can ask for work again.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I have no problems downloading, it's uploads that are stuck.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

I'd love to know the trick to clear the stalled uploads. Nothing I have tried has allowed them to upload. Several thousand uploads pending on each host.

Not sure about "thousands", but I just use the "transfers" tab in Boinc Tasks and a lot of mouse clicks!!
Saying that, all my Linux machines have cleared their backlog, however both Windows machines with less uploads take forever.

Clicking Update does nothing since each machine has a constant 8 tasks Active with all the rest pending. When a task does eventually upload, it sits at 100% and then waits from a minute to a half hour before getting an ACK.
I run spoofed clients, so "thousands" is accurate and not some exaggeration. The Threadripper host has 3300 tasks in Upload currently.

. . OK, so you're the culprit ... :)

Stephen

:)

All clear here on my 6 linux computers with GPUs. One has an upload that hasn't cleared and its back off is now 5 hours. It is strange because when a task finishes that same computer uploads it in about 20 seconds. This one upload won't upload even when I hit retry. At least I am back to crunching Seti.

Try suspending network communication in the Manager, wait a minute and restart network communication, then hit Retry. That sometimes unsticks a recalcitrant upload.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

It looks like uploads have started to resume.

---

Yes, I have finally managed to clear my uploads. Now starting to get downloads but some of them are stalling out. So still a struggle to rebuild caches.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Well I've suspended GPU processing on my 2 rigs until I can get a bit of a buffer downloaded, but it's not fun downloading all this work when most of it is just garbage. :-(

Well all that crap that I downloaded has exceeded my peak internet allowance so I'm on a slow boat now until the end of the month during the day. :-(

Cheers.

That really sucks ...

And now the splitters have fallen on their face with the downloads finally working. RTS basically nil.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)