Lifemapper work units

Brucifer

Member
Oct 31, 2004
48
0
0

Starting late yesterday and continuing through the night and also this morning, I have been getting close to 50% of my downloaded work unit files with a length of 1. Interspersed are also the 160 byte file lengths, however these are outnumbered now by the 1-byte files. The problem is that it is taking much longer to download a load of 50 work units for my linux boxes. Question: Is anyone else noticing this issue? I didn't have this problem before the latest database issues yesterday.
 

Unforgiven

Golden Member
May 11, 2001
1,827
0
0
same thing here man, you arent alone. started happening yesterday for me as well. about 50% 1 k or less :(
 

Brucifer

Member
Oct 31, 2004
48
0
0

Talk about a fast answer! :) Well I'm glad someone else is having the problem *ONLY* because it at least lets me know that my delivery leg isn't the cuplrit.

Also, there should be something posted today on Spectre's lifemapper forum regarding the future plans and direction of the project. At least according to what was posted by the staff member a week ago after all the comments/complaints aobut the total lack of communication and response to problems. Not that I really expect to see anything there mind you, but I thought I'd play along for another week with their game.

For me it has been a hassle keeping work units on hand to force feed my systems to keep the production up during this push to take overall position #2 on the Lifemapper project. This latest issue with the small corrupted work units has caused me to lose ground. My LM target was two things, 1) for the team to reach #2, and 2) for me to reach 100,000. I am getitng very close to reaching my 2nd goal, as we now have the #2 position overall. So I will most likely be dropping my LM output significantly, and switching over to another project that runs hands-off very reliably, like seti classic, TSC/D2OL, RC5, etc. :)

Hopefully the LM staff will get their "stuff" together and cure some of their more pressing issues as I do like the project.
 

Wolfsraider

Diamond Member
Jan 27, 2002
8,305
0
76
I have been babysitting here too
I am running on linux as well

but its actually starting to get better.

yesterday thisproblem was really bad
 

Unforgiven

Golden Member
May 11, 2001
1,827
0
0
well brucifer i can completely understand your frustrations. you saw me post on the planet spectre forums in telling the admins that their communication blows and i have seen it a lot since ive joined the project. seti classic is a stable project and if you connect to a seti queue you are almost never down! i love lifemapper and will continue to contribute but i will indeed have backup projects installed on all machines i have that will run seti classic when lifemapper is having issues. in fact, having a backup project is always a good thing :)
 

Brucifer

Member
Oct 31, 2004
48
0
0

With seti classic (and before when it was just seti) I use(d) Seti-Queue. A good little add-on that works like a charm. Too bad that there isn't something like that for LM. Of course seti keeps their work units in the same format.... so you can queue stuff up. Whereas LM changes stuff around and has said in the past that they re-do their stuff on a whim, so that if you hang on to stuff for 3 days or more that it may not be the current thing, but that you'll still get credit for it but the completed units won't add to the current effort.

Backup stuff... yep, I keep some TSC units loaded into clients so that if my unstable internet connection goes on the fritz, I always have something for the boxes to grind on. It would have been nice if LM would have increased the queue on linux LM from 40 to 200 like windows. Each system goes through 2 50 unit loads in 24 hours. If I leave the stuff on autopilot, then the conenction problems start up, and the sum of it all is that I end up with idle systems, so that's why I just download a wad of 50 unit batches onto one system, and then use that as a rotating pool for my other systems to sftp into to put the completed stuff and get new stuff. The overall time loss for tics involved in sending/receiving and putting up with timedout result uploads being cached for later insertion is much less. Then my "pivot" system is just an older box that otherwise wouldn't have the pleasure of serving the LM effort. :)