Lost my block for no apparent reason

[rsbriggs]

What's up with this?

[Sun Jul 06 06:18:39 2003] connecting to server
[Sun Jul 06 06:18:43 2003] logging into server
[Sun Jul 06 06:18:43 2003] requesting a block
[Sun Jul 06 06:18:47 2003] got proth test from server (k=5359, n=4047126)
[Sun Jul 06 06:29:23 2003] resolving hostname
[Sun Jul 06 06:29:26 2003] opening connection
[Sun Jul 06 06:29:26 2003] receiving from server
[Sun Jul 06 06:29:27 2003] logging into server
[Sun Jul 06 06:29:27 2003] login successful
[Sun Jul 06 06:29:28 2003] n.high = 15263 . 265 blocks left in test
...
[Mon Jul 07 20:15:13 2003] n.high = 747887 . 217 blocks left in test
[Mon Jul 07 23:26:27 2003] resolving hostname
[Mon Jul 07 23:26:30 2003] opening connection
[Mon Jul 07 23:26:30 2003] receiving from server
[Mon Jul 07 23:26:31 2003] logging into server
[Mon Jul 07 23:26:31 2003] login successful
[Mon Jul 07 23:26:32 2003] n.high = 763150 . 216 blocks left in test
[Mon Jul 07 23:26:35 2003] resolving hostname
[Mon Jul 07 23:27:39 2003] opening connection
[Mon Jul 07 23:27:40 2003] receiving from server
[Mon Jul 07 23:27:42 2003] logging into server
[Mon Jul 07 23:27:42 2003] requesting a block
[Mon Jul 07 23:27:47 2003] got proth test from server (k=27653, n=4059897)

:cry:

[Roman]

I also would like to ask, what is this? The log is from a quite slow machine, but I thought if it reports back at least every 5 days, the test won't be abandoned.

[Thu Aug 21 03:17:54 2003] iteration: 2160000/4231610 (51.04%) k = 55459 n = 4231594
[Thu Aug 21 03:43:46 2003] logging into server
[Thu Aug 21 03:43:47 2003] login successful
[Thu Aug 21 03:43:48 2003] n.high = 2163955 . 149 blocks left in test
[Thu Aug 21 15:15:24 2003] client process [v1.0.2] invoked
[Thu Aug 21 15:15:25 2003] priority set to idle
[Thu Aug 21 15:15:25 2003] got k and n from cache
[Thu Aug 21 15:15:26 2003] restarting proth test from cache (k=55459, n=4231594) [51.1%]
[Thu Aug 21 15:20:09 2003] temporarily unable to connect -- block added to submit queue
[Thu Aug 21 15:59:28 2003] iteration: 2170000/4231610 (51.28%) k = 55459 n = 4231594
[Thu Aug 21 16:50:58 2003] temporarily unable to connect -- block added to submit queue
[Thu Aug 21 17:04:31 2003] iteration: 2180000/4231610 (51.52%) k = 55459 n = 4231594
[Thu Aug 21 18:09:38 2003] iteration: 2190000/4231610 (51.75%) k = 55459 n = 4231594
[Thu Aug 21 18:21:52 2003] temporarily unable to connect -- block added to submit queue
[Thu Aug 21 19:14:41 2003] iteration: 2200000/4231610 (51.99%) k = 55459 n = 4231594
[Thu Aug 21 19:52:40 2003] logging into server
[Thu Aug 21 19:52:41 2003] login successful
[Thu Aug 21 19:52:42 2003] n.high = 2205838 . 146 blocks left in test
[Thu Aug 21 20:19:46 2003] iteration: 2210000/4231610 (52.23%) k = 55459 n = 4231594
[Thu Aug 21 21:25:32 2003] temporarily unable to connect -- block added to submit queue
[Thu Aug 21 21:27:19 2003] iteration: 2220000/4231610 (52.46%) k = 55459 n = 4231594
[Thu Aug 21 22:33:37 2003] iteration: 2230000/4231610 (52.70%) k = 55459 n = 4231594
[Thu Aug 21 22:58:07 2003] logging into server
[Thu Aug 21 22:58:08 2003] login successful
[Thu Aug 21 22:58:08 2003] n.high = 2233760 . 144 blocks left in test
[Thu Aug 21 23:38:43 2003] iteration: 2240000/4231610 (52.93%) k = 55459 n = 4231594
[Fri Aug 22 00:29:21 2003] logging into server
[Fri Aug 22 00:29:24 2003] login successful
[Fri Aug 22 00:31:04 2003] server had no record of proth test, test abandoned
[Fri Aug 22 00:31:04 2003] connecting to server
[Fri Aug 22 00:31:05 2003] couldn't get block from server [can't connect], retrying in 120 secs
[Fri Aug 22 00:33:05 2003] connecting to server
[Fri Aug 22 00:33:47 2003] couldn't get block from server [can't connect], retrying in 120 secs
[Fri Aug 22 00:35:47 2003] connecting to server
[Fri Aug 22 00:35:55 2003] logging into server
[Fri Aug 22 00:36:00 2003] requesting a block
[Fri Aug 22 00:36:09 2003] got proth test from server (k=21181, n=4377860)
[Fri Aug 22 00:36:09 2003] server packet cached to disk
[Fri Aug 22 01:42:04 2003] iteration: 10000/4377875 (0.23%) k = 21181 n = 4377860
[Fri Aug 22 02:03:56 2003] logging into server
[Fri Aug 22 02:03:57 2003] login successful
[Fri Aug 22 02:03:58 2003] n.high = 13044 . 335 blocks left in test
[Fri Aug 22 02:50:38 2003] iteration: 20000/4377875 (0.46%) k = 21181 n = 4377860
[Fri Aug 22 03:31:11 2003] temporarily unable to connect -- block added to submit queue

[srowen]

I also just experienced a similar problem -- the client abandoned my test because "server had no record of proth test", though it had been running and sending in updates every 30 minutes or so. See below. Note that this abandoned test still shows up in my "Pending Test" screen.

Is the test abandoned if 10 days have elapsed since the client started on the test, or since the client's last report to the server? It's the latter, right? Then I'm confused about what happened here...


...
[Fri Aug 29 10:29:29 2003] logging into server
[Fri Aug 29 10:29:29 2003] login successful
[Fri Aug 29 10:29:30 2003] n.high = 2999776 . 81 blocks left in test
[Fri Aug 29 10:29:58 2003] iteration: 3000000/4163435 (72.06%) k = 5359 n = 416
3422
[Fri Aug 29 10:51:03 2003] iteration: 3010000/4163435 (72.30%) k = 5359 n = 416
3422
[Fri Aug 29 10:59:55 2003] logging into server
[Fri Aug 29 10:59:55 2003] login successful
[Fri Aug 29 11:00:16 2003] server had no record of proth test, test abandoned
[Fri Aug 29 11:00:16 2003] connecting to server
[Fri Aug 29 11:00:16 2003] logging into server
[Fri Aug 29 11:00:16 2003] requesting a block
[Fri Aug 29 11:00:25 2003] got proth test from server (k=19249, n=4434026)
[Fri Aug 29 11:00:25 2003] server packet cached to disk
[Fri Aug 29 11:21:23 2003] iteration: 10000/4434041 (0.23%) k = 19249 n = 44340
26
[Fri Aug 29 11:27:05 2003] logging into server
[Fri Aug 29 11:27:05 2003] login successful
[Fri Aug 29 11:27:05 2003] n.high = 12715 . 348 blocks left in test

[srowen]

I forgot to mention a few details about my setup:

Client version 1.0.2
Redhat Linux 9.0
AMD Duron 750MHz
256MB RAM

I started processing the block that was lost on about August 1 (the computer is on every few days) -- I'm not sure if that has anything to do with it.

[itsjustme]

After seeing your post, I checked the log on my linux machine :

grep "no record" sclient.log

and found the no record of proth problem had occurred 3 times since August 22nd. Up until the lost record problem, submittals were occuring every 15 minutes or so without problems.

Anyone else seeing this happen?

I am running:

Linux Kernel 2.4.21
client v1.0.2
Athlon 1700+
512 Megs Ram


Example Record:

[Fri Aug 22 17:53:32 2003] n.high = 2600928 . 135 blocks left in test
[Fri Aug 22 18:02:57 2003] iteration: 2610000/4362539 (59.83%) k = 21181 n = 4362524
[Fri Aug 22 18:07:14 2003] logging into server
[Fri Aug 22 18:07:15 2003] login successful
[Fri Aug 22 18:07:35 2003] server had no record of proth test, test abandoned
[Fri Aug 22 18:07:35 2003] connecting to server
[Fri Aug 22 18:07:36 2003] logging into server
[Fri Aug 22 18:07:36 2003] requesting a block
[Fri Aug 22 18:07:40 2003] got proth test from server (k=4847, n=4274391)
[Fri Aug 22 18:07:40 2003] server packet cached to disk
[Fri Aug 22 18:18:11 2003] iteration: 10000/4274404 (0.23%) k = 4847 n = 4274391

[wirthi]

Got a similar problem. I loose the current block every time I finish quit the client (yes, I have clicked on "stop" prior shuting the client down ...)

Some time ago I've read that "you have to reboot after installing" the client. That's what I have done ...

Is that a known problem or just happening on my machine(s)?

[srowen]

That might be a different issue... there is no GUI for the Linux client and no Stop button or anything.

[wirthi]

Originally posted by srowen
That might be a different issue... there is no GUI for the Linux client and no Stop button or anything.
That's a good point. Actually I was talking about the Windows (GUI) version. I thought that was what the original poster (rsbriggs) was talking about ...

[dfamily]

here's one from the for what it's worth dept...

you are assigned a test but if for some reason there is a problem
getting it, on the next try you may also get a new and different
test...note the the following - it doesn't look strange at all but...

Sep 02 21:23:37 2003] logging into server
[Tue Sep 02 21:23:37 2003] login successful
[Tue Sep 02 21:23:37 2003] n.high = 4440150 . 1 blocks left in test
[Tue Sep 02 21:28:51 2003] residue: C531EADA78A484CB
[Tue Sep 02 21:28:51 2003] completed proth test(k=21181, n=4445492): result 3
[Tue Sep 02 21:28:51 2003] connecting to server
[Tue Sep 02 21:28:53 2003] logging into server
[Tue Sep 02 21:28:53 2003] requesting a block
[Tue Sep 02 21:29:04 2003] couldn't get block from server [block denied], retry in 180 secs
[Tue Sep 02 21:32:04 2003] connecting to server
[Tue Sep 02 21:32:14 2003] logging into server
[Tue Sep 02 21:32:18 2003] requesting a block
[Tue Sep 02 21:32:21 2003] got proth test from server (k=24737, n=4467103)
[Tue Sep 02 21:44:33 2003] resolving hostname
[Tue Sep 02 21:44:33 2003] opening connection
[Tue Sep 02 21:44:33 2003] receiving from server
[Tue Sep 02 21:44:34 2003] logging into server
[Tue Sep 02 21:44:34 2003] login successful
[Tue Sep 02 21:44:34 2003] n.high = 12528 . 356 blocks left in test

what you don't see is that according to my pending test list, I had
been assigned 24737-4394527 at the 21:29 mark before the
block denied line - and in this particular case, didn't even get the
zfile (is the zfile something we get or do we create it on our
end?)...anywho, someone else has alluded to if there's trouble
getting a block it can create interesting happenings and I can
now concur - it's happened before but I couldn't see why...now it
makes sense...

patrick.

[Mystwalker]

The z file is something that's created by the client. You only get the k/n-pair from the server.

[rsbriggs]

This has happened twice again in the last couple of days. Needed to re-boot, paused the client, rebooted, and lost the current block in progress:

[Sun Sep 07 01:24:36 2003] n.high = 320788 . 339 blocks left in test
[Sun Sep 07 08:48:38 2003] block processing paused
[Sun Sep 07 09:39:15 2003] block processing resumed
[Sun Sep 07 14:27:14 2003] resolving hostname
[Sun Sep 07 14:27:19 2003] opening connection
[Sun Sep 07 14:27:19 2003] receiving from server
[Sun Sep 07 14:27:20 2003] logging into server
[Sun Sep 07 14:27:20 2003] login successful
[Sun Sep 07 14:27:20 2003] n.high = 333126 . 338 blocks left in test
[Sun Sep 07 19:02:16 2003] block processing paused
(reboot here)
[Sun Sep 07 19:05:49 2003] connecting to server
[Sun Sep 07 19:06:02 2003] logging into server
[Sun Sep 07 19:06:02 2003] requesting a block
[Sun Sep 07 19:06:12 2003] got proth test from server (k=28433, n=4509265)

[Ixlandia]

After doing some regular maintenance on my machines and reboots, I just lost all the current progress of all the blocks I was testing on my Windows machines even though at shutdown of the client it said progress saved. Really? Bad, really bad. After restarting of the client it requested a new block and started from scratch on a new one.

After losing days of work for this project on mulitple machines due to the inability of the client to save progress reliably, I have removed SB off my machines and went back to GIMP's which has a much better track record of reliably saving work in progress. This is not an isolated incident and is something I have experienced with SB in the past but this time will be the last.

Now I see why in the one article written about this project, they claim it may take decades to complete this project. No wonder. If everyone has to keep restarting tests from scratch after machine crashes, reboots, whatever, and with an inability to restart tests from where the client left off, then indeed I see why it would take so long to complete the project. Losing work as a result of a simple exiting of the client without a reliable save function certainly is not going to make the project complete any faster if you always have to start from scratch on a new block. Good luck!

[MathGuy]

We seem to have two threads here discussing what *may* be the same problem. How many of you who have experienced a lost block are using the windows client with the service handler and the KeepVisible option?

The discussion in this thread:

http://www.free-dc.org/forum/showthread.php?s=&threadid=3268

gives a possible solution (if that's what the problem is...)

[allio]

The difference, I think, is that some people are actually getting assigned a new block halfway through their test, and others are simply restarting their block without getting a new one.

[Dærk]

Hey how do I get my old block back?:swear:

[wirthi]

Originally posted by Dærk
Hey how do I get my old block back?:swear:
Good question; if you have a backup of both the necessary files (some looking like z1234566.*) and the registry keys (when using windows) it should be possible to restore a test. Anybody tried that?

[Mystwalker]

The registry keys can be set easily without having backed up earlier - you just need to look up the SB log file to get the k/n-pair.
But I'm not sure whether or not the server still accepts blocks for that test...

[MikeH]

But I'm not sure whether or not the server still accepts blocks for that test... The server will accept the blocks and the result (and you'll get credit in the stats), but there are two points worth noting.

1. If the candidate isn't in your pending test list 'preferences->pending test management', then it will have been assigned to someone else already, so you''ll be duplicating work.

2. You'll end up on Louie's "suspicious" list (which you don't need to worry about if you're not doing anything naughty :whip: ).

[kulp]

Regarding getting new blocks when previous ones aren't complete, apparently it matters what your current directory is when you run sb. In other words, when i start it once from its own directory, it'll cache there, but if i give its full path from another dir., it'll cache in the directory i'm in. So just cd to a certain directory every time and you'll be fine.

It was a bit annoying till i found this out.

Actually i just noticed this was suggested (on the Windows version by jjjl at http://www.free-dc.org/forum/showthread.php?s=&threadid=3268

--kulp