32 hour WU on one machine won't upload. One file upload is at 100%, the other is at 86.454%

From the log:

yafu 2/23/2018 9:24:00 AM Started upload of yafu_ali_119640_L142_C128_1519179608_20_0_r151467157_0
yafu 2/23/2018 9:24:00 AM Started upload of yafu_ali_203040_L142_C129_1519246810_58_0_r1117734217_0
yafu 2/23/2018 9:24:02 AM Temporarily failed upload of yafu_ali_119640_L142_C128_1519179608_20_0_r151467157_0: transient upload error
yafu 2/23/2018 9:24:02 AM Backing off 03:21:33 on upload of yafu_ali_119640_L142_C128_1519179608_20_0_r151467157_0
yafu 2/23/2018 9:24:05 AM Temporarily failed upload of yafu_ali_203040_L142_C129_1519246810_58_0_r1117734217_0: transient upload error
yafu 2/23/2018 9:24:05 AM Backing off 04:14:38 on upload of yafu_ali_203040_L142_C129_1519246810_58_0_r1117734217_0
yafu 2/23/2018 9:24:05 AM [error] No scheduler URLs found in master file

Also this message:

yafu 2/23/2018 9:24:05 AM [error] No scheduler URLs found in master file

Here's the WU (looks like 2 WU's?):
yafu 134.05 yafu-4t (4t) 4C 00:00:09 (00:00:00) yafu_ali_203040_L142_C129_1519246810_58_0 2/23/2018 7:12:36 AM 0.00 Uploading
yafu 134.05 yafu-4t (4t) 4C 32:24:27 (15:07:33) yafu_ali_119640_L142_C128_1519179608_20_0 2/23/2018 7:12:04 AM 46.67 Uploading


I suspect this has something to do with the server outage? What to do?

I said heck with it and just deleted the project and tried to reattach. It says "Project temporarily unavailable" when I try to attach. Yet all the site pages seem to be online.

BTW, can't seem to find a URL listed on the main page or instructions for attaching to YAFU anywhere. The URL I've used before is: http://yafu.myfirewall.org/yafu/

The url http://yafu.myfirewall.org/yafu/ should be fine. I think you see it on the page only if you are not logged in.

OK, logging out allows one to wade through a few screens to get to the URL. However there doesn't seem to be any way to get there if you're logged in and are simply adding a machine.

The more important thing is why the project keeps reporting "temporarily unavailable" when trying to add a machine?

I detached a system and attached it new to yafu without problems.

The URL I use to attach: http://yafu.myfirewall.org/yafu/

When I try to attach via BOINC manager I get "project unavailable".

When I try to attach via BoincTasks it attempts to attach, says the project is available then in the BOINC log:

http://yafu.myfirewall.org/yafu/ 2/23/2018 3:48:28 PM [error] No scheduler URLs found in master file

The same message as caused the upload message in my first post above.

Installed an older version of BOINC (7.6.33), still happens. I'll try running out all the other work from other projects (all working fine), completely uninstall BOINC and blow its directories away and reinstall. In all the many years that I've run BOINC I've never seen this one...

This seemed to have coincided with the server outage, maybe a coincidence?

What happens on a different computer?

What happens on a different computer?

Removed YAFU on a different computer and it reattached fine.

On the problem computer I just uninstalled BOINC, deleted all BOINC directories (both program and data), even searched the registry for YAFU references. Reinstalled BOINC 7.8.3, installed other projects. All that went perfectly. Then tried to install YAFU. "The project is temporarily unavailable, please try again later."

Here's the host in question:

https://yafu.myfirewall.org/yafu/show_host_detail.php?hostid=31175

Two of the three tasks that show as in progress are the ones that partially uploaded and then couldn't finish and report. The third I aborted before originally deleting the project after the upload problem. None of this seemed to be acknowledged by the server. Very strange. Nothing more that I can do at this end.

Edit: My other 10 YAFU machines are running fine.

On server side they look fine.

Any firewall / antivirus settings?

There was some option in Boinc to switch to http 1.0. Try this.

Disabled AV and firewall: no change
Changed computer name and rebooted: no change
Added <http_1_0>1</http_1_0> to the config and checked event log to be sure it was enabled: no change

My theory is that this machine was in the process of transferring files to the server during the outage and something changed.
There is nothing left of the previous BOINC client on the machine yet YAFU and only YAFU reports as "temporarily unavailable".
Mysterious. I wonder if it will work after those 3 WUs that the sever thinks are in process time out?

Maybe we need to call on Hercule Poirot?

All your hosts connecting to yafu with the same IP.
So, it is not that yafu blocks the IP. In the case no host would be able to connect.

Can you post the message log when you press update?

You mean when I try to add the project? This is the only line added in the log:

2/25/2018 2:10:50 PM |  | Fetching configuration file from http://yafu.myfirewall.org/yafu/get_project_config.php

It looks like there is no response from the server.


If I update on a different machine, I "almost instantly" get:

2/25/2018 2:15:44 PM | yafu | update requested by user
2/25/2018 2:15:48 PM | yafu | Sending scheduler request: Requested by user.
2/25/2018 2:15:48 PM | yafu | Requesting new tasks for CPU
2/25/2018 2:15:51 PM | yafu | Scheduler request completed: got 0 new tasks
2/25/2018 2:15:51 PM | yafu | No tasks sent
2/25/2018 2:15:51 PM | yafu | No tasks are available for YAFU-4t
2/25/2018 2:15:51 PM | yafu | No tasks are available for YAFU-8t
2/25/2018 2:15:51 PM | yafu | This computer has reached a limit on tasks in progress

The other 10 machines are all running YAFU very nicely. The only thing to mention is that quite a few of the 4t and 8t WUs with shorter deadlines are ending up running at "high priority" in BOINC and even over the deadline. None of the WUs are ever swapped out, as the "switch between app time" is set at 80 hours. Basically YAFU runs all the time on these computers, only running other apps as filler when there are spare cores.

My theory is that this machine was in the process of transferring files to the server during the outage and something changed.
There is nothing left of the previous BOINC client on the machine yet YAFU and only YAFU reports as "temporarily unavailable".
Mysterious. I wonder if it will work after those 3 WUs that the sever thinks are in process time out?

After the 3 WUs mentioned above timed out the server allowed the client to reattach normally as its original ID.

Oh boy, thought this was fixed. After successfully completing 10 WUs on this machine it's happening again. From the BOINC log:

43 yafu 3/8/2018 11:05:42 AM Sending scheduler request: Requested by project.
44 yafu 3/8/2018 11:05:42 AM Requesting new tasks for CPU and NVIDIA GPU
45 yafu 3/8/2018 11:05:48 AM [error] No start tag in scheduler reply

68 yafu 3/8/2018 11:07:24 AM Fetching scheduler list
69 yafu 3/8/2018 11:07:25 AM [error] No scheduler URLs found in master file

211 yafu 3/8/2018 11:48:17 AM Started upload of yafu_ali_1140402_L137_C127_1520399441_14_0_r1191318454_0
212 yafu 3/8/2018 11:48:17 AM Started upload of yafu_ali_1240372_L130_C125_1520361615_32_1_r1743751288_0
213 yafu 3/8/2018 11:48:18 AM Temporarily failed upload of yafu_ali_1140402_L137_C127_1520399441_14_0_r1191318454_0: transient upload error
214 yafu 3/8/2018 11:48:18 AM Backing off 04:14:26 on upload of yafu_ali_1140402_L137_C127_1520399441_14_0_r1191318454_0
215 yafu 3/8/2018 11:48:18 AM Temporarily failed upload of yafu_ali_1240372_L130_C125_1520361615_32_1_r1743751288_0: transient upload error
216 yafu 3/8/2018 11:48:18 AM Backing off 02:40:10 on upload of yafu_ali_1240372_L130_C125_1520361615_32_1_r1743751288_0

The machine (the two in progress WUs are actually completed but won't finish uploading): https://yafu.myfirewall.org/yafu/results.php?hostid=31175
Another 26 hours of WUs that won't complete uploading. One is a stuck upload at 100%, the other at 83.9%.
Rebooting the computer didn't help. All other projects (as before) are working perfectly. I give up.

A follow up. Left the 2 WUs trying to upload while gone for 9 days. Since the website records disappear quickly I don't know if they finally uploaded after they timed out (doubt it), but after the time out date the computer has again received WUs and the completed WUs are now once again being accepted by the server. Will continue to observe...

Mysterious.

Not very.
While outage, local Computer got wrong entry into its local DNS cache.
Flushing DNS would have helped - waiting did same Thing as local DNS tables have timestamps.

entering ipconfig /flushdns into a command window would have done it.