MD5 mismatch - Unexpected error during p110825543.proof.tmp upload:

[drkirkby]

I was, up to a few minutes ago, performing PRP tests on 110825543 and 33264623 with mprime Linux64,v30.3,build 6. I am using a Dell 7920 with 2 x Intel Xeons 8167M with 320 GB of error correcting RAM. The operating system is CentOS 7.9. The latter exponent, which is only 47.48% complete and has 12 days to go. (That's the first and last time I attempt a 100 million digit exponent). However, the former

https://www.mersenne.org/report_expo...0825543&full=1

has just completed, but with an error message.

[Comm thread Apr 18 08:52] Unexpected error during p110825543.proof.tmp  upload:  {"error_status":401,"error_message":"Unauthorized","error_description":"JSON  result not yet submitted by your user ID or proof file MD5 mismatch"}

(See a bigger section of the log I keep below. The assignment was allocated to me, and is shown as completed, but not prime. I chanced my username on here, as I think that's supposed to be private, although personally I have no problems with anyone knowing it.

[Worker #1 Apr 18 08:39] Iteration: 110670000 / 110825543 [99.85%], ms/iter:  1.594, ETA: 00:04:07
[Worker #1 Apr 18 08:40] Iteration: 110680000 / 110825543 [99.86%], ms/iter:  1.595, ETA: 00:03:52
[Worker #1 Apr 18 08:40] Iteration: 110690000 / 110825543 [99.87%], ms/iter:  1.593, ETA: 00:03:35
[Worker #1 Apr 18 08:40] Iteration: 110700000 / 110825543 [99.88%], ms/iter:  1.593, ETA: 00:03:20
[Worker #2 Apr 18 08:40] Iteration: 157960000 / 332646233 [47.48%], ms/iter:  6.283, ETA: 12d 16:52
[Worker #1 Apr 18 08:40] Iteration: 110710000 / 110825543 [99.89%], ms/iter:  1.593, ETA: 00:03:04
[Worker #1 Apr 18 08:41] Iteration: 110720000 / 110825543 [99.90%], ms/iter:  1.595, ETA: 00:02:48
[Worker #1 Apr 18 08:41] Iteration: 110730000 / 110825543 [99.91%], ms/iter:  1.593, ETA: 00:02:32
[Worker #1 Apr 18 08:41] Iteration: 110740000 / 110825543 [99.92%], ms/iter:  1.594, ETA: 00:02:16
[Worker #2 Apr 18 08:41] Iteration: 157970000 / 332646233 [47.48%], ms/iter:  6.282, ETA: 12d 16:47
[Worker #1 Apr 18 08:41] Iteration: 110750000 / 110825543 [99.93%], ms/iter:  1.594, ETA: 00:02:00
[Worker #1 Apr 18 08:42] Iteration: 110760000 / 110825543 [99.94%], ms/iter:  1.596, ETA: 00:01:44
[Worker #1 Apr 18 08:42] Iteration: 110770000 / 110825543 [99.94%], ms/iter:  1.593, ETA: 00:01:28
[Worker #1 Apr 18 08:42] Iteration: 110780000 / 110825543 [99.95%], ms/iter:  1.593, ETA: 00:01:12
[Worker #2 Apr 18 08:42] Iteration: 157980000 / 332646233 [47.49%], ms/iter:  6.282, ETA: 12d 16:47
[Worker #1 Apr 18 08:43] Iteration: 110790000 / 110825543 [99.96%], ms/iter:  1.593, ETA: 00:00:56
[Worker #1 Apr 18 08:43] Iteration: 110800000 / 110825543 [99.97%], ms/iter:  1.594, ETA: 00:00:40
[Worker #1 Apr 18 08:43] Iteration: 110810000 / 110825543 [99.98%], ms/iter:  1.593, ETA: 00:00:24
[Worker #1 Apr 18 08:43] Iteration: 110820000 / 110825543 [99.99%], ms/iter:  1.593, ETA: 00:00:08
[Worker #2 Apr 18 08:43] Iteration: 157990000 / 332646233 [47.49%], ms/iter:  6.282, ETA: 12d 16:47
[Worker #1 Apr 18 08:44] Gerbicz error check passed at iteration 110824464.
[Worker #1 Apr 18 08:44] Gerbicz error check passed at iteration 110825488.
[Worker #1 Apr 18 08:44] Gerbicz error check passed at iteration 110825537.
[Worker #1 Apr 18 08:44] Generating proof for M110825543.  Proof power = 10, Hash length = 64
[Worker #1 Apr 18 08:44] Root hash = EEE6DA14BDEE9361F4A6E6062790F74358220040AE72E49309A597F4EF86FAE7
[Worker #1 Apr 18 08:44] hash0 = 6E14AF90BC1BDA4B
[Worker #1 Apr 18 08:44] hash1 = 2EC6CEEC6EA64231
[Worker #2 Apr 18 08:44] Iteration: 158000000 / 332646233 [47.49%], ms/iter:  6.271, ETA: 12d 16:14
[Worker #1 Apr 18 08:45] hash2 = 63774B8586EF3235
[Worker #1 Apr 18 08:45] hash3 = 5C1DFE612B985A33
[Worker #2 Apr 18 08:45] Gerbicz error check passed at iteration 158000000.
[Worker #1 Apr 18 08:45] hash4 = 1BFCAE0BBE325F2D
[Worker #1 Apr 18 08:45] hash5 = 97D95FE3786A464B
[Worker #2 Apr 18 08:46] Iteration: 158010000 / 332646233 [47.50%], ms/iter:  6.278, ETA: 12d 16:32
[Worker #1 Apr 18 08:46] hash6 = 4885CF2192025660
[Worker #2 Apr 18 08:47] Iteration: 158020000 / 332646233 [47.50%], ms/iter:  6.272, ETA: 12d 16:14
[Worker #2 Apr 18 08:48] Iteration: 158030000 / 332646233 [47.50%], ms/iter:  6.265, ETA: 12d 15:52
[Worker #1 Apr 18 08:48] hash7 = D0850EFEB0A00DCD
[Worker #2 Apr 18 08:49] Iteration: 158040000 / 332646233 [47.50%], ms/iter:  6.265, ETA: 12d 15:52
[Worker #2 Apr 18 08:50] Iteration: 158050000 / 332646233 [47.51%], ms/iter:  6.264, ETA: 12d 15:47
[Worker #2 Apr 18 08:51] Iteration: 158060000 / 332646233 [47.51%], ms/iter:  6.261, ETA: 12d 15:38
[Worker #1 Apr 18 08:52] hash8 = C7CE4B9C2508B6C6
[Comm thread Apr 18 08:52] MD5 of p110825543.proof.tmp is 93b8faea6ac412ad3e91a058ae5244a0
[Comm thread Apr 18 08:52] Proof file exponent is 110825543
[Comm thread Apr 18 08:52] Filesize of p110825543.proof.tmp is 138531989
[Comm thread Apr 18 08:52] URL: http://www.mersenne.org/proof_upload/?UserID=foo&Exponent=110825543&FileSize=138531989&FileMD5=93b8faea6ac412ad3e91a058ae5244a0
[Comm thread Apr 18 08:52] == Info:   Trying 162.212.57.131:80...
[Comm thread Apr 18 08:52] == Info: Connected to www.mersenne.org (162.212.57.131) port 80 (#0)
[Comm thread Apr 18 08:52] => Send header: GET /proof_upload/?UserID=xxxxxxx&Exponent=110825543&FileSize=138531989&FileMD5=93b8faea6ac412ad3e91a058ae5244a0 HTTP/1.1
[Comm thread Apr 18 08:52] Host: www.mersenne.org
[Comm thread Apr 18 08:52] Accept: */*
[Comm thread Apr 18 08:52]
[Comm thread Apr 18 08:52] == Info: Mark bundle as not supporting multiuse
[Comm thread Apr 18 08:52] <= Recv header: HTTP/1.1 401 Unauthorized 97
[Comm thread Apr 18 08:52] <= Recv header: Cache-Control: no-store, no-cache, must-revalidate
[Comm thread Apr 18 08:52] <= Recv header: Pragma: no-cache
[Comm thread Apr 18 08:52] <= Recv header: Content-Type: application/json
[Comm thread Apr 18 08:52] <= Recv header: Expires: Thu, 19 Nov 1981 08:52:00 GMT
[Comm thread Apr 18 08:52] <= Recv header: Server: Microsoft-IIS/8.5
[Comm thread Apr 18 08:52] <= Recv header: X-Powered-By: PHP/7.1.10
[Comm thread Apr 18 08:52] <= Recv header: Set-Cookie: GIMPSWWW=n4q4ds4e69vcnrcs2nijd4inn7; expires=Tue, 18-May-2021 07:52:03 GMT; Max-Age=2592000; path=/; domain=mersenne.org
[Comm thread Apr 18 08:52] <= Recv header: Date: Sun, 18 Apr 2021 07:52:02 GMT
[Comm thread Apr 18 08:52] <= Recv header: Content-Length: 146
[Comm thread Apr 18 08:52] <= Recv header:
[Comm thread Apr 18 08:52] <= Recv data: {"error_status":401,"error_message":"Unauthorized","error_description":"JSON result not yet submitted by your user ID or proof file MD5 mismatch"}
[Comm thread Apr 18 08:52] == Info: Connection #0 to host www.mersenne.org left intact
[Comm thread Apr 18 08:52] Unexpected error during p110825543.proof.tmp upload: {"error_status":401,"error_message":"Unauthorized","error_description":"JSON result not yet submitted by your user ID or proof file MD5 mismatch"}
[Worker #2 Apr 18 08:52] Iteration: 158070000 / 332646233 [47.51%], ms/iter:  6.267, ETA: 1

[drkirkby]

The certificate eventually got uploaded, and the result verified by someone using the certificate.

https://www.mersenne.org/report_expo...0825543&full=1
The only problem I noticed is that there are two 14 MB files left over
m110825543 m110825543.bu
I will just delete them. So the problem has been resolved.

[Prime95]

Apparently someone at Microsoft thought it would be really cool if FindFirstFile returns xxx.proof.tmp when asking for all file names matching *.proof

[slandrum]

Originally Posted by Prime95

Apparently someone at Microsoft thought it would be really cool if FindFirstFile returns xxx.proof.tmp when asking for all file names matching *.proof

In the Tomb of the Unknown Implementor...

Upon entering the tomb, one is greeted by a hollow voice that says, "That's not a bug, it's a feature!"

[drkirkby]

Originally Posted by Prime95

Apparently someone at Microsoft thought it would be really cool if FindFirstFile returns xxx.proof.tmp when asking for all file names matching *.proof

As stated in the original post, I was running CentOS 7.9 - i.e. Linux, so as much as I do like to bash Microsoft over Windoze, it's difficult to blame M$ here.

[Prime95]

Originally Posted by drkirkby

As stated in the original post, I was running CentOS 7.9 - i.e. Linux, so as much as I do like to bash Microsoft over Windoze, it's difficult to blame M$ here.

Might have to bash me, off to check the Linux code....

[Prime95]

Originally Posted by Prime95

Might have to bash me, off to check the Linux code....

Yup, mea culpa, as Chris likes to say "stupid programmer error".

[drkirkby]

Originally Posted by Prime95

Might have to bash me, off to check the Linux code....

Okay, consider yourself bashed to look at the linux code. Could I suggest two more things that could do with improving.

1) When it says "uploading binary data" or something like that, it would be good to know what is being uploaded. I see some messages later in the day about uploading the binary data, but did not know if they were the proof file or not.

2) I can set the Daytime P-1/ECM stage 2 memory to 100 GB, and the Nighttime P-1/ECM stage 2 memory to 100 GB, However, I'm limited to setting the maximum emergency limit to a lower figure of 78.582520 GB which does not make any sense to me. How can the emergency limit be lower the non-emergency limit?

Daytime P-1/ECM stage 2 memory in GB (100.000000):
Nighttime P-1/ECM stage 2 memory in GB (100.000000):
Max emergency memory in GB/worker (70.000000): 120
Please enter a value between 0.000000 and 78.582520.

[kruoli]

Regarding 2)
These are memory settings for totally different things. The first two (as indicated) are only used in stage 2 of P-1 and ECM. The third one is memory reserved for temporary unavailable file writes, i.e. if you save your proof files to a network share and it's disconnected in the moment the proof gets written, it will be stored in the emergency memory until the network share is reconnected.

[drkirkby]

Originally Posted by kruoli

Regarding 2)
These are memory settings for totally different things. The first two (as indicated) are only used in stage 2 of P-1 and ECM. The third one is memory reserved for temporary unavailable file writes, i.e. if you save your proof files to a network share and it's disconnected in the moment the proof gets written, it will be stored in the emergency memory until the network share is reconnected.

Thank you.

[AwesomeMachine]

Am I too understand the problem in the OP is now fixed? I just have to download the Linux software again?