Message boards : Server and website : Problem with upload server?
Author | Message |
---|---|
For a few hours, I've been experiencing a problem with uploading finished tasks: the upload speed which is normally between 500 and 600 KBps, now varies between 20 KBps and almost zero. The upload of about 50% of a Gianni WU (181MB size) has been running for almost 80 minutes now. | |
ID: 44323 | Rating: 0 | rate: / Reply Quote | |
The uploads and downloads are probably being slowed by the bandwidth of the Adria OPEN and CLOSED tasks downloading, failing right away, then reporting failed. Bad task batches currently. Need to just wait this out until they are removed again. | |
ID: 44340 | Rating: 0 | rate: / Reply Quote | |
The uploads and downloads are probably being slowed by the bandwidth of the Adria OPEN and CLOSED tasks downloading, failing right away, then reporting failed. Bad task batches currently. Need to just wait this out until they are removed again. this might be the explanation. I, too, am receiving these Adria WUs on all of my hosts; no idea what they are not being removed from the server. | |
ID: 44341 | Rating: 0 | rate: / Reply Quote | |
since this afternoon, upload speed is back to normal. | |
ID: 44347 | Rating: 0 | rate: / Reply Quote | |
Uploading a work unit, 145MiB, speed varies between 1mbit/s and 10 kilobit/sec. $ ping -qc 50 84.89.134.145 --- 84.89.134.145 ping statistics --- 50 packets transmitted, 50 received, 0% packet loss, time 49056ms rtt min/avg/max/mdev = 370.399/371.614/373.334/0.616 ms Is there a problem currently at the server side, or could it be something else I have overlooked. Are others having troubles uploading WUs? | |
ID: 47851 | Rating: 0 | rate: / Reply Quote | |
again, I see extremely slow uploads today; many times, they even pause for a while. | |
ID: 47883 | Rating: 0 | rate: / Reply Quote | |
I have given up my hope to ever get an answer from the GPUGRID people here :-( | |
ID: 48030 | Rating: 0 | rate: / Reply Quote | |
I have given up my hope to ever get an answer from the GPUGRID people here :-( I have not noticed slow uploads. But even if they are slow, as long as the upload doesn't take days, it doesn't really affect the flow. As long as your GPU is not idling, I don't see a problem. Also, I don't have a problem with the GPUGrid Scientists not answering every single comment because I'd rather them concentrate on the entire reason they are doing this in the first place. To make groundbreaking scientific discoveries, and if they are answering questions on the forums every day they can't possibly be doing that. They are low on people at the moment and don't even currently have an IT professional helping them so unfortunately we have to deal with this decrepit system at the moment. | |
ID: 48033 | Rating: 0 | rate: / Reply Quote | |
Also, I don't have a problem with the GPUGrid Scientists not answering every single comment Well, fact though is that not a single comment is answered. None at all :-( And this can be quite frustrating for volunteers after a while. | |
ID: 48035 | Rating: 0 | rate: / Reply Quote | |
I work with scientists every day and I can tell you time is not something they have spare of. They are in meetings most of the day discussing the science at hand and when they're not in meetings they should be working. I think they should keep and eye on the server status and allow the work to always be flowing but for them to read every comment and respond is just not feasible. | |
ID: 48036 | Rating: 0 | rate: / Reply Quote | |
They could appoint other students as PR representatives and surely the university has access to IT students and staff. | |
ID: 48037 | Rating: 0 | rate: / Reply Quote | |
As I had already written in another thread here in the forum some time ago (after having read this as a comment of one of the project administrators in the forum of another BOINC project): | |
ID: 48038 | Rating: 0 | rate: / Reply Quote | |
I remember that as well Erich and yet nothing happens. Only leaves me to think that those were empty words. | |
ID: 48039 | Rating: 0 | rate: / Reply Quote | |
Do we have WUs to crunch? We do. Are there new publications proving that our crunching is useful? There are. | |
ID: 48040 | Rating: 0 | rate: / Reply Quote | |
Another uncertainty, I am afraid, now is what will happen to Universitat Pompeu Fabra in Barcelona in view of the forthcoming political developments/conflicts between Madrid and Catalonia. | |
ID: 48041 | Rating: 0 | rate: / Reply Quote | |
Since last night, the upload server again is extremely slow. The upload speed of finished tasks is a fraction of what it normally is, sometimes it even comes to a halt. Uploads take forever. | |
ID: 48298 | Rating: 0 | rate: / Reply Quote | |
Since last night, the upload server again is extremely slow. The upload speed of finished tasks is a fraction of what it normally is, sometimes it even comes to a halt. Uploads take forever.I don't have any upload problems since last night. Usually it's only you reporting this slowness, so perhaps the error is on your ISP's side. Try http://www.speedtest.net/; Open it with Microsoft Edge (or any flash enabled browser), and then you can select the location of the server you want to test your connection with. Practically you should select one in Barcelona. If the upload speed to those servers is slow too, then it's sure that GPUGrid has nothing to do about this. | |
ID: 48299 | Rating: 0 | rate: / Reply Quote | |
I know this thread is about uploads, but right now I have an ADRIA_FOLDCALP80 download that has periods of less than 1KBps download speed. Most files downloaded normally, but two took much longer. The last file (...pdb_file) is at 4.7% and 17 minutes.... | |
ID: 48313 | Rating: 0 | rate: / Reply Quote | |
I don't have any upload problems since last night. Usually it's only you reporting this slowness, so perhaps the error is on your ISP's side. Thanks, Zoltan, for your hint regarding speedtest.net. So far, I've used this tool for testing my own internet connection when some problem came up once in a while. Following your advice, from December 8th until today, I have tested the connection from here to servers in Barcelona and also the area around Barcelona numerous times. EVERYTIME the results were between 125 and 145 Mbps for download, and between 13,5 and 15 Mbps for upload. Which would, in theory, support my assumption that the problem of the sometimes very low upload speed has to do with the GPUGRID upload server. Unless there are other possibilities, technical-wise, which I am not aware of (not being a network expert at all). I think that the fact that I am the only one who reports these flaws here in the Forum does not neccessarily mean that I am the only one with this problem. I guess many other crunchers simply may not notice it, or if they notice it, they do not care. On the other hand, you, for example, did NOT have such problems. No idea what's really behind this, and how it could be found out and solved. To show the actual figures: on my WinXP machines, the upload speed to GPUGRID normally is between 280 and 380 KBps. On my Windows10 machines it normally is between 600 and 700 KBps. Whenever the problem comes up, above cited speeds are reduced by 80-90%. | |
ID: 48334 | Rating: 0 | rate: / Reply Quote | |
What I've noticed about uploads to GPUGrid are that 150kB is a good upload speed. Many times I've seen about 90kB. | |
ID: 48340 | Rating: 0 | rate: / Reply Quote | |
That sounds like the expected behaviour if the communications channel - end-to-end, anywhere along the line - has an intermediate device which is dropping packets. | |
ID: 48349 | Rating: 0 | rate: / Reply Quote | |
I've just watched task 16759934 uploading from the north of England. 107 MB, showing ~445 KB/sec steady speed, completed in 246 seconds (at that speed, it should have been 240 seconds, so a little loss). | |
ID: 48351 | Rating: 0 | rate: / Reply Quote | |
since yesterday, the upload speed is back to what it normally is: on my WinXP machines, the upload speed to GPUGRID normally is between 280 and 380 KBps. so let's see for how long this will last. | |
ID: 48352 | Rating: 0 | rate: / Reply Quote | |
Hi all, traceroute -I www.gpugrid.org
traceroute to www.gpugrid.org (84.89.134.145), 30 hops max, 60 byte packets
4 87.186.200.142 (87.186.200.142) 25.197 ms 25.216 ms 25.416 ms
5 f-ed8-i.F.DE.NET.DTAG.DE (217.5.95.114) 27.937 ms 28.805 ms 28.823 ms
6 te0-0-0-5-4.agr41.fra03.atlas.cogentco.com (130.117.14.149) 28.825 ms 24.536 ms 24.509 ms
7 be3186.ccr41.fra03.atlas.cogentco.com (130.117.0.1) 24.495 ms 24.475 ms 24.454 ms
8 be2799.ccr41.par01.atlas.cogentco.com (154.54.58.234) 37.299 ms 37.221 ms 37.146 ms
9 be3325.ccr21.bio02.atlas.cogentco.com (130.117.48.205) 49.148 ms 49.379 ms 49.283 ms
10 be3357.ccr31.mad05.atlas.cogentco.com (130.117.1.21) 54.259 ms 54.088 ms 54.006 ms
11 be3375.agr22.mad05.atlas.cogentco.com (130.117.50.202) 55.501 ms 55.724 ms 55.645 ms
12 be2853.rcr11.b015537-1.mad05.atlas.cogentco.com (154.54.56.62) 51.609 ms 53.262 ms 53.184 ms
13 149.11.68.2 (149.11.68.2) 56.547 ms 55.245 ms 55.159 ms
14 CIEMAT.AE2.telmad.rt4.mad.red.rediris.es (130.206.245.2) 52.929 ms 52.273 ms 52.207 ms
15 TELMAD.AE4.uv.rt1.val.red.rediris.es (130.206.245.89) 58.730 ms 58.265 ms 58.181 ms
16 anella-val1-router.red.rediris.es (130.206.211.70) 66.893 ms 68.186 ms *
17 * * *
18 84.89.159.147 (84.89.159.147) 63.540 ms 63.841 ms 63.673 ms
19 * * *
20 * * *
21 grosso.upf.edu (84.89.134.145) 67.743 ms 68.738 ms 68.571 ms Cheers | |
ID: 48355 | Rating: 0 | rate: / Reply Quote | |
104MB Upload this morning looked fine until about 70%, then: | |
ID: 48367 | Rating: 0 | rate: / Reply Quote | |
Based on Richard Haselgrove's comments I suggest to post some traceroute paths. Mine looks like this (using Windows): tracert www.gpugrid.org
Tracing route to www.gpugrid.org [84.89.134.145]
over a maximum of 30 hops:
5 13 ms 13 ms 12 ms core1-hu0-17-0-5.colindale.ukcore.bt.net [195.99.127.150]
6 14 ms 14 ms 14 ms peer6-hu0-6-0-2.telehouse.ukcore.bt.net [62.172.103.13]
7 13 ms 13 ms 12 ms gi3-0.lonth-inter-1.interoute.net [195.66.224.53]
8 28 ms 28 ms 26 ms ae1-0.par-gar-score-1-re0.interoute.net [212.23.42.21]
9 34 ms 35 ms 34 ms ae1-0.mad-001-score-1-re0.interoute.net [89.202.161.21]
10 35 ms 35 ms 34 ms 206.pool1.dialup.cybernet-ag.de [195.143.124.206]
11 42 ms 42 ms 41 ms TELMAD.AE4.uv.rt1.val.red.rediris.es [130.206.245.89]
12 * 56 ms 56 ms anella-val1-router.red.rediris.es [130.206.211.70]
13 * * * Request timed out.
14 * 55 ms 55 ms 84.89.159.147
15 * * * Request timed out.
16 * * * Request timed out.
17 48 ms 49 ms 47 ms grosso.upf.edu [84.89.134.145] Why on earth do I go via a dialup pool in Germany? | |
ID: 48369 | Rating: 0 | rate: / Reply Quote | |
Here's my tracert from Nebraska, USA. Looks like I wander in atlas.congentco] system for quite a distance, but have the same timeouts near grosso.upf.edu, including that device with no DNS entry. | |
ID: 48372 | Rating: 0 | rate: / Reply Quote | |
From Northern Virginia VZ FiOS. 4 10 ms 6 ms 7 ms B3361.WASHDC-LCR-21.verizon-gni.net [100.41.128.16] 5 * * * Request timed out. 6 10 ms 8 ms 9 ms 0.et-5-1-0.BR2.IAD8.ALTER.NET [140.222.239.85] 7 13 ms 12 ms 12 ms be3018.ccr41.iad02.atlas.cogentco.com [154.54.10.197] 8 12 ms 11 ms 9 ms be3084.ccr42.dca01.atlas.cogentco.com [154.54.30.65] 9 16 ms 17 ms 15 ms be2807.ccr42.jfk02.atlas.cogentco.com [154.54.40.109] 10 82 ms 81 ms 82 ms be2490.ccr42.lon13.atlas.cogentco.com [154.54.42.86] 11 92 ms 92 ms 93 ms be12489.ccr42.par01.atlas.cogentco.com [154.54.57.70] 12 104 ms 104 ms 105 ms be3324.ccr52.bio02.atlas.cogentco.com [130.117.2.65] 13 104 ms 106 ms 106 ms be3358.ccr32.mad05.atlas.cogentco.com [130.117.1.97] 14 106 ms 106 ms 106 ms be3378.agr21.mad05.atlas.cogentco.com [154.54.36.174] 15 105 ms 107 ms 106 ms be2852.rcr11.b015537-1.mad05.atlas.cogentco.com [154.54.36.166] 16 112 ms 113 ms 112 ms 149.11.68.2 17 113 ms 113 ms 115 ms CIEMAT.AE2.telmad.rt4.mad.red.rediris.es [130.206.245.2] 18 123 ms 124 ms 122 ms TELMAD.AE4.uv.rt1.val.red.rediris.es [130.206.245.89] 19 123 ms 118 ms 118 ms anella-val1-router.red.rediris.es [130.206.211.70] 20 * * * Request timed out. 21 127 ms 126 ms 124 ms 84.89.159.147 22 * * * Request timed out. 23 * * * Request timed out. 24 138 ms 142 ms 140 ms grosso.upf.edu [84.89.134.145] And using a VPN with a Swiss location. 1 118 ms 118 ms 118 ms 10.69.10.1 2 118 ms 119 ms 120 ms 5.157.7.81 3 120 ms 121 ms 119 ms 5.157.4.25 4 119 ms 120 ms 119 ms 83.231.187.177 5 140 ms 139 ms 138 ms ae-16.r24.frnkge08.de.bb.gin.ntt.net [129.250.7.16] 6 152 ms 151 ms 150 ms ae-5.r24.londen12.uk.bb.gin.ntt.net [129.250.3.12] 7 238 ms 171 ms 175 ms ae-6.r01.mdrdsp03.es.bb.gin.ntt.net [129.250.4.138] 8 182 ms 174 ms 174 ms rediris.baja.espanix.net [193.149.1.26] 9 185 ms 186 ms 185 ms CIEMAT.AE1.cica.rt1.and.red.rediris.es [130.206.245.38] 10 200 ms 200 ms 203 ms CICA.AE1.uv.rt1.val.red.rediris.es [130.206.245.34] 11 206 ms 206 ms 207 ms anella-val1-router.red.rediris.es [130.206.211.70] 12 * * * Request timed out. 13 204 ms 203 ms 202 ms 84.89.159.147 14 * * * Request timed out. 15 * * * Request timed out. 16 * 205 ms 206 ms grosso.upf.edu [84.89.134.145] | |
ID: 48451 | Rating: 0 | rate: / Reply Quote | |
half an hour ago, a finished task got uploaded with "normal" speed; a few minutes later, the next one goes with about a tenth of the "normal" speed. | |
ID: 48493 | Rating: 0 | rate: / Reply Quote | |
"Request Timed Out" is not necessarily a problem or fault as there is a lot of hardware on the internet which doesn't respond to pings. | |
ID: 48494 | Rating: 0 | rate: / Reply Quote | |
Again, big upload problems all day long. Now the uploads even come to a halt and change to "retry" status. | |
ID: 48738 | Rating: 0 | rate: / Reply Quote | |
The couple times I have caught an upload recently it had been much improved. Sometimes over 1gig. It still had the jump up in upload speed, a gradual fall off and jump back up but from 150m to 1g is a nice jump. | |
ID: 48739 | Rating: 0 | rate: / Reply Quote | |
again, upload is extremely slow (this has also been the case several time within the past days) - about 20 KB/s (normally, it's between 250 and 350 KB/s). | |
ID: 49229 | Rating: 0 | rate: / Reply Quote | |
today, once more major problems with the upload server. Uploads take forever. | |
ID: 49681 | Rating: 0 | rate: / Reply Quote | |
I am unable to reach the GPUGrid website (forums, etc.) several times a day. It is annoying when you are trying to change settings, or post a comment. But since the uploads retry often enough, it is not really a problem for that. | |
ID: 49682 | Rating: 0 | rate: / Reply Quote | |
But since the uploads retry often enough, it is not really a problem for that.except that when the crunching time before was - say - 23 hours (which is often the case with my two GTX750ti's), and the upload takes, as happened today, 1 1/2 hours. Thus you loose 20% of the credit points. I am surprised there are not more comments about it however.so am I | |
ID: 49684 | Rating: 0 | rate: / Reply Quote | |
except that when the crunching time before was - say - 23 hours (which is often the case with my two GTX750ti's), and the upload takes, as happened today, 1 1/2 hours. Thus you loose 20% of the credit points. You have reminded me that I have a GTX 750 Ti sitting on the shelf, and I just converted a machine to Ubuntu 16.04, where I think it would get in under the 24 hour limit with the new Linux app. It is still about the most efficient card out there, and should get me through the summer, since I have had to shut down my GTX 970 and 980. It should work. You should try Ubuntu. It is awful to learn, but once you get it then it is actually faster to install than Windows now. | |
ID: 49687 | Rating: 0 | rate: / Reply Quote | |
You have reminded me that I have a GTX 750 Ti sitting on the shelf, and I just converted a machine to Ubuntu 16.04, where I think it would get in under the 24 hour limit with the new Linux app. My first Pablo_IDP came it at 10 hours, and Adria_FOLDT0986 came in at 10 1/2 hours, so it works. http://www.gpugrid.net/results.php?hostid=433866 For a card with a TDP of 60 watts, that will do. EDIT: The Adria was about what I would expect, based on my GTX 970 and 980, but the Pablo was unusually fast. I would expect about 20 to 22 hours. We will see. | |
ID: 49692 | Rating: 0 | rate: / Reply Quote | |
[quote]My first Pablo_IDP came it at 10 hours, and Adria_FOLDT0986 came in at 10 1/2 hours, so it works. :-) :-) :-) | |
ID: 49696 | Rating: 0 | rate: / Reply Quote | |
I knew you would like it. | |
ID: 49697 | Rating: 0 | rate: / Reply Quote | |
Message boards : Server and website : Problem with upload server?