[06:54:20] + Attempting to send results [September 5 06:54:20 UTC]
[06:54:41] - Couldn't send HTTP request to server
[06:54:41] + Could not connect to Work Server (results)
[06:54:41] (130.237.232.141:8080)
[06:54:41] + Retrying using alternative port
[06:55:02] - Couldn't send HTTP request to server
[06:55:02] + Could not connect to Work Server (results)
[06:55:02] (130.237.232.141:80)
[06:55:02] - Error: Could not transmit unit 07 (completed September 4) to work server.
[06:55:02] + Attempting to send results [September 5 06:55:02 UTC]
[06:55:02] - Couldn't send HTTP request to server
[06:55:02] + Could not connect to Work Server (results)
[06:55:02] (130.237.165.141:8080)
[06:55:02] + Retrying using alternative port
[06:55:02] - Couldn't send HTTP request to server
[06:55:02] + Could not connect to Work Server (results)
[06:55:02] (130.237.165.141:80)
[06:55:02] Could not transmit unit 07 to Collection server; keeping in queue.
[06:55:02] - Failed to send all units to server
Folding@Home Client Shutdown.
--- Opening Log file [September 5 06:55:52 UTC]
# Windows SMP Console Edition #################################################
###############################################################################
Folding@Home Client Version 6.34
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: C:\Program Files (x86)\SMP
Executable: C:\Program Files (x86)\SMP\FAH6.34-win32-SMP.exe
.
Why wasn't/isn't the CS collecting Work Results that can't be accepted by the WS (concerned) ?
.
- stopped Linux SMP w. HT on i7-860@3.5 GHz
....................................
Folded since 10-06-04 till 09-2010
Sorry about that. The client should try to resend it soon and it should go. Apparently that server went down recently, but they took care of the problem and put it online again. See: viewtopic.php?f=18&t=19522
EDIT: if you are in a hurry you could add the "-send all" flag and restart the client and it should resend that WU.
EDIT 2: oh looks like some are still having trouble: viewtopic.php?f=18&t=19508&start=30#p194719
F@h is now the top computing platform on the planet and nothing unites people like a dedicated fight against a common enemy. This virus affects all of us. Lets end it together.
It was the reason why a lot of Results from the .141 WU's could not be uploaded; the .141 server was down (offline) and the CS would not accept them either.
This isn't supposed to happen and it is disastrous (for F@H) in a 3 day holiday weekend !
More so for someone running a big Folding farm.
- stopped Linux SMP w. HT on i7-860@3.5 GHz
....................................
Folded since 10-06-04 till 09-2010
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
CS servers haven't worked well since client version 5. I've seen several statements in that regard in this forum. I'm not sure why you would expect otherwise at this point.
However, I have also seen where there are updates planned for the collection servers, in conjuction with the V7 client, to make the collection servers work much better.
Jesse_V wrote:if you are in a hurry you could add the "-send all" flag and restart the client and it should resend that WU.
That's entirely unnecssary. The statement should read: "If you are in a hurry, restart the client and it should resent that WU."
When you restart without the -send all flag, the client resumes work on the current WU and attempts to send any WUs which are queued.
When you restart with the -send all flag, the client does NOT resumes work on the current WU and attempts to send any WUs which are queued before shutting down.
In other words, the -send all flag has nothing to do with sending queued work. All it does is tell the client to stop after it tries rather than continuing to process the current WU.
From what is displayed in the Server Stats it seems that none of the CS' are accepting at this time; that explains it.
Now the question is why?
3 Collection Servers and not one doing what it was set up to do ... http://fah-web.stanford.edu/pybeta/serverstat.html
.
- stopped Linux SMP w. HT on i7-860@3.5 GHz
....................................
Folded since 10-06-04 till 09-2010
same problem, losing points all the time that I can't upload. Their going to get slammed when they finally get back online too as there going to be an awful lot of units trying to send in.
At least the AS are working so we can get more work.
7im's earlier statements are entirely correct. Let me fill in a bit more detail, though.
The old Collection Server code has been broken for a long time and there are several reasons why it couldn't be corrected quickly. A new version is being developed and tested partly as an adjunct to the needs of the V7 client development. Rest assured that the Pande Group is just as interested in getting work returned quickly as you are and I they've been aware of all of the issues you raise and have been/are working actively to resolve them.
You do realize that nobody notices how long the CSs have been off-line until there are failures of a couple of important Work Servers like we had yesterday. The CS code is only needed when Work Servers fail (which is rare) so the big job of getting the CS code working again doesn't rate the same priority as any number of other developmental projects which are in constant use. Nevertheless, the plan to fix the CSs has been quietly progressing with the development of some radically new code.
As you might expect, any bug in the server code carries a very high risk since discovering it and fixing it might affect many people in a very short time. Developing and testing new server code (whether it's CS or WS or AS or Stats server) is necessarily a slow process that must be managed very, very carefully. As with all future developments, it will be put into production whenever it's ready without any announcement of a "planned date." (Since this is mostly a behind-the-scenes effort, I really doubt that there will even be a statement about "soon" vs. "not soon")
I was clearly talking about CS4, 5 and 6, the new guys on the block; the old CS has indeed been out for ages. We 're not counting on that one.
Donors are counting on 1 or any of the new group of 3 Collection Servers to help them and F@H out when 1 or more Work Servers go to 'Not Accept' or 'Down', etc..
At least one of the three current servers should be 'Online' at all time, otherwise they have no reason for being, at all ...
By the way; they currently are (still) 'Not Accept' and 'standby', all three of them!
.
- stopped Linux SMP w. HT on i7-860@3.5 GHz
....................................
Folded since 10-06-04 till 09-2010
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
At least one of the three current servers should be 'Online' at all time, otherwise they have no reason for being, at all ...
Not even to test the new CS software? All or nothing seems a little bit extreme to me...
Noorman, did you read my previous comment? Testing of real-time changes such as new CSs must be done VERY, VERY carefully. 7im is correct when he points out that the new CSs are still being validated. It's unrealistic to leave a new CS on-line unmonitored when running unrelased code, even if the code is a release-candidate. The old code is completely dead and the new code is still being validated. Almost all uploads have to go to the production work servers and the backup function that the CSs may soon provide still needs to be considered unavailable, no matter what you see on serverstat.
Why then not let the old CS 'working' in parallel?
Now, there 's just nothing (anymore), no backup at all for the accepting of Results created by dedicated donors who don't want to see work lost or delayed ...
.
- stopped Linux SMP w. HT on i7-860@3.5 GHz
....................................
Folded since 10-06-04 till 09-2010
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
Has anyone noted success with the Windows 6.34 client automatically switching back to the a5 client from the a3? In the past I've always done this manually. Right now my machine is churning through a3 SMP WUs. Is there a way to know when the bigadv WUs AR available again? Thanks.
[Edit] Yes is the answer. Automatically issued an a5 WU 2685.
Sat Sep 10 22:29:05
