Very interesting test assignment

[vjs]

I noticed one of my computers downloaded two test at one time??? This computer has been crunching for a while without a problem until now.

Do these two k/n pair magically match or something??

335233 21181•2^6272492+1 129.7.239.89 Mon May 31 19:57:49 2004 Thu Jun 3 16:40:41 2004 1080180 17 % 171801


335236 27653•2^6272529+1 129.7.239.89 Mon May 31 20:02:53 2004 Thu Jun 3 16:48:03 2004 1080180 17 % 171707

Notice they are also progressing together???

[kugano]

Do you possibly have two machines on the same IP, e.g. with NAT? If not, could you post the logfile from that client since the test(s) was downloaded?

Server log for the first test:

% grep "k = 21181, n = 6272492" server.log | head -n 5
2004-05-31 15:57:49.060 INFO  [DataManager] Retrieved assignment k = 21181, n = 6272492 for user 5202 (requested k = 0)
2004-05-31 16:16:54.722 INFO  [DataManager] Updated progress: at bit = 6354 for k = 21181, n = 6272492 by user 5202
2004-05-31 16:41:07.666 INFO  [DataManager] Updated progress: at bit = 12708 for k = 21181, n = 6272492 by user 5202
2004-05-31 17:05:19.558 INFO  [DataManager] Updated progress: at bit = 19062 for k = 21181, n = 6272492 by user 5202
2004-05-31 17:29:32.468 INFO  [DataManager] Updated progress: at bit = 25416 for k = 21181, n = 6272492 by user 5202

... and the other one:

% grep "k = 27653, n = 6272529" server.log | head -n 5
2004-05-31 16:02:53.701 INFO  [DataManager] Retrieved assignment k = 27653, n = 6272529 for user 5202 (requested k = 0)
2004-05-31 16:27:06.265 INFO  [DataManager] Updated progress: at bit = 6354 for k = 27653, n = 6272529 by user 5202
2004-05-31 16:51:16.836 INFO  [DataManager] Updated progress: at bit = 12708 for k = 27653, n = 6272529 by user 5202
2004-05-31 17:15:28.634 INFO  [DataManager] Updated progress: at bit = 19062 for k = 27653, n = 6272529 by user 5202
2004-05-31 17:39:39.116 INFO  [DataManager] Updated progress: at bit = 25416 for k = 27653, n = 6272529 by user 5202

So it looks like there really are two "independent" sets of reports coming in...

[MikeH]

Are you using the service handler? Maybe it has created two instances of the client (for some reason). This would at least make sense of the two tests running at the same rate.

Are these rates roughly half what you have seen (for a single instance) in the past.

[vjs]

I was just in this weekend checking on the borged machine.

Looks like I have a colaborator, I've installed the client as a service and I or someone else has started it with the start menu.

I'll check the reg for cache and cache2 on monday.

It seemed weird that they were at the same progress level etc.

I'll also check the log file to see if anything looks fishy.

[kugano]

It seemed weird that they were at the same progress level etc.

The two tests had very close n values, and it's the n value that determines how often a test gets reported. Apparently they were close enough to cause the first few reports, at least, to happen at the same iteration (but at different time intervals).