Corrupt WU's

[Martyn]

Man I'm so racked off. I've been running nonet for a few days, no reason particularly, just do sometimes. I decided to dump and I'd no cached WU's. I must've got a corrupt WU or something - or a few of em. I'd been noticing the client crashing a bit lately, but just re-started and continued. I did seem to notice that nearly every time I booted into windows, the client would load and try to phone home for a WU. It registered with me, but I've been so preoccupied lately, I didn't investigate. Dammit, i've got 0 production for nearly a week! I've got a few WU's on a second machine, but they're 23aa's and woth squat. Is anyone else having this problem?

[Stilgar]

Kind of strange a while back I picked up a few for a short time that seemed to have problems. Had about 4 on different machines what went belly up. All with in a couple days. But none recently.

[wylie]

Ive had more bad units than usual recently. Usually at night....probably just after I turn the screen off & go to bed...:mad:

[wylie]

Originally posted by tril0Byte
I did seem to notice that nearly every time I booted into windows, the client would load and try to phone home for a WU. It registered with me, but I've been so preoccupied lately,[/B]
Are you saying your client cant connect & get a unit?
If so I have 2 machines like that. I just shared the g@h directory, went to another computer, read the directory in network neighbourhood (dont map the drive....for some reason it doesnt work when the drive is mapped), ran genome where it sent its cache and got fresh werk. didnt even need to stop the client.
Hope this helps.

[siggy]

Seems back before the .98 came out, I was getting bad WU regularly. But now (knock on wood) I have not had any in a while. Course on the other hand I am also not getting any small WU's. I am seeing 99aa, 98aa and 88aa. Whats with that?

[Stilgar]

Originally posted by siggy
Seems back before the .98 came out, I was getting bad WU regularly. But now (knock on wood) I have not had any in a while. Course on the other hand I am also not getting any small WU's. I am seeing 99aa, 98aa and 88aa. Whats with that?

Luck of the draw. They have added several different types lately.

[phil]

I picked up a 12aa the other day on my 1575Mhz TBird which I -nonetted over night just to see what happened (it was my first small one :D) 17 Genes later, I uploaded. I have started to check what each machine is doing lately and swap work units around. It seems daft the TBird doing a 12aa and my CII600@1035 doing a 99aa.

[zhotfire]

I lost 3 of 4 genes on the weekend. Sequences 22, 23 and 28... all between 78-92aa.... :( The fourth crashed in the 29th sequence but was able to restart....

[dnar]

The worse case I have experienced was during the last Drag Race. 3 bad units in 4 days....:mad:

My last bad unit crashed as I watched, just before my big dump on the weekend :D Close.

[siggy]

I heard there is a new weighing system. Anyone know how it works?

[Dave S]

Hummmm....

Had 6 last week, but this week so far so good!!:)

[siggy]

Course once you have a good one just go -nonet and forget it for a while.

[pelligrini]

Originally posted by tril0Byte
Man I'm so racked off. I've been running nonet for a few days, no reason particularly, just do sometimes. I decided to dump and I'd no cached WU's. I must've got a corrupt WU or something - or a few of em. I'd been noticing the client crashing a bit lately, but just re-started and continued. I did seem to notice that nearly every time I booted into windows, the client would load and try to phone home for a WU. It registered with me, but I've been so preoccupied lately, I didn't investigate. Dammit, i've got 0 production for nearly a week! I've got a few WU's on a second machine, but they're 23aa's and woth squat. Is anyone else having this problem?

As I understand it, it's a bug in the client, not corrupt WU's. I believe it's a combination of certain seed numbers and the AA sequences that cause the client to crash.

I have had my fair share of crashes, but my machines at the office running -nonet have a much, much higher failure rate than all of my cable connected machines at home. It is unusual to find all of the clients running after a couple of days. I had a 75% failure rate over last weekend. I don't have any evidence to back it up, but I think the seed numbers generated by the clients while in -nonet exacerbates the problem. I haven’t had a client crash at home this month, while I had one machine crash a 99aa at seq 30 (real pisser) and another crunch a 26aa 4 times before crashing just Monday night. I hope Stephan tracked down the bug (which he said is there) while he’s on vacation.

I do end up going in to the office more often now, just to baby-sit some clients :(

[zhotfire]

Originally posted by siggy
Course once you have a good one just go -nonet and forget it for a while.
I suppose you'd do that if you needed to, but remember the woonits dished out by the Stanford server are timely. They need to be crunched along with the other units given out so that the research can continue. There's nothing worse than having to wait for results when you have analysis to do..... :(
:crunchndump:

[Dave S]

Originally posted by pelligrini


As I understand it, it's a bug in the client, not corrupt WU's. ...........and another crunch a 26aa 4 times before crashing just Monday night. :(

arr.......yep had the same thing. crunched one 4 or 5 times before it crashed!!:(

[Dave S]

Originally posted by Dave S


arr.......yep had the same thing. crunched one 4 or 5 times before it crashed!!:(

Just had another one!! DOH!!! crunched it twice, third time round @10 it crashes the client...dam:( - had to dump it and get another one :confused:

[verT]

I hope you upload those genes that are completed first before you go clearing them

[dnar]

Originally posted by verT

I hope you upload those genes that are completed first before you go clearing them



The last time I had a bad whoonit, I went "-clear" before sending around 6 completed genes back - ARGGGGHHHH! But guess what???? clearing a bad unit does not delete the ones that were completed successfully.....:D

[Dave S]

Originally posted by verT
I hope you upload those genes that are completed first before you go clearing them

OYer:D:D

[verT]

The only reason I asked was because the first time I got a bad WU running nonet I had two or three finished and it pooped out so I cleared bad woonit and retarted......... oops :rolleyes: um I don't do that any more....

[dnar]

What are toy guys on about????? As I have stated in the post above, I have tried clearing the current bad unit and the completed units still remain.....:confused:

[verT]

I only did this once and it wiped everything out for me.

[dnar]

Originally posted by verT

I only did this once and it wiped everything out for me.



Maybe the 0.98 Linux client has the edge...:D

[Dave S]

Originally posted by dnar
What are toy guys on about????? As I have stated in the post above, I have tried clearing the current bad unit and the completed units still remain.....:confused:
:confused: worked ok 4 me in 0.98 & W2K

[verT]

It did happen on the first ver of 0.98 maybe that was it.... I've since upgraded to all later ver. of 0.98. Hasn't happened since so I guess that was it. Or maybe there is a spy somewhere? Hmmmm, I do work with my brother and he is a moron......

[Dave S]

Originally posted by verT
It did happen on the first ver of 0.98 maybe that was it.... I've since upgraded to all later ver. of 0.98. Hasn't happened since so I guess that was it. Or maybe there is a spy somewhere? Hmmmm, I do work with my brother and he is a moron......

Yer could have been the first version, strange tho :confused:
A lesion learned is a WU gained;)

[MechCD]

Whats with my 1.33ghz tbird getting 3-4 99aa units and my 560mhz P!! getting 6-7 12aa units and then 2 99aa units? Wouldn't it be better just to give the 560 a bunch of 60-75aa units instead of the blasted 12aas?

My 1.33 ghz once got a 12aa and it went by sooooo fast i almost stopped the client and did a -clear ;) (but i didn't)

[siggy]

Originally posted by dnar




Maybe the 0.98 Linux client has the edge...:D

Maybe it is a linux thing? No that's not possible Linux is perfect.

[Dave S]

Hummm just had 3 in a row, all in the 80 range:confused:

[siggy]

Dang. I got a bad WU sometime in the night so I ended up not cruching at all.

I need to get an alarm on my computer when a WU is bad.

[stall6g]

had 2 bad work units this morning. One was a 92 that was almost near the end of completing sequence 30 and the other was an 82aa at sequence 23. Why can't they crash on the little ones? That way at least I wouldn't lose so much when I have to clear the bad work.:(

[LBaker]

I'm probably going to curse myself but I've been very lucky. I've only had 3 or 4 bad units since I started genoming. They were bigger ones and almost done before they crashed.

[siggy]

I wish they would crash while it is filtering. obviously it is not filtering everything.

[dnar]

I run my clients as background tasks, with a bunch of Linux scripts taking care of the monitoring and execution there of...

If a client dies, it is restarted within 60 seconds. Next modification will be to automate tje clearing of a bad whoonit if it crashes more than once.... Anyone interested in the code?

[Dave S]

Originally posted by dnar
I run my clients as background tasks, with a bunch of Linux scripts taking care of the monitoring and execution there of...

If a client dies, it is restarted within 60 seconds. Next modification will be to automate tje clearing of a bad whoonit if it crashes more than once.... Anyone interested in the code?
Hummmm I guess if I can run these as a “service” in W2K I could do something similar, can it be done :confused: has anyone done this??

[LBaker]

Originally posted by LBaker
I'm probably going to curse myself but I've been very lucky. I've only had 3 or 4 bad units since I started genoming. They were bigger ones and almost done before they crashed.

I knew I should have kept my mouth shut :mad: I just got a bad one. Luckily it only went to sequence 4 before crashing!

[Dave S]

Originally posted by LBaker


I knew I should have kept my mouth shut :mad: I just got a bad one. Luckily it only went to sequence 4 before crashing!
lol take a look at this:confused:
http://wasya.com/images/pc1gh0.jpg

[LBaker]

What the H is that. Are they all corrupt units?

[Dave S]

Originally posted by LBaker
What the H is that. Are they all corrupt units?
Nope don’t think so, it recovers & continues ok:confused:

[Stilgar]

Originally posted by LBaker
What the H is that. Are they all corrupt units?

It looks like he was lucky and it corrected itself. Had a few of those myself. They seem to make it thru though.

[LBaker]

Originally posted by Dave S

Nope don’t think so, it recovers & continues ok:confused:

That's GOOD news.:D

[siggy]

I had another bad one during the night. I have never had one correct itself, that I know of.

[Dave S]

Originally posted by siggy
I had another bad one during the night. I have never had one correct itself, that I know of.
Bad luck girl:(

[siggy]

Thanks I have a good one running now I think.

[Dave S]

Originally posted by siggy
Thanks I have a good running now I think.
good:)

[siggy]

Yep it was a good one. I should have kept it an -noneted for a while. Oh well maybe next time.

[wylie]

bloody 28th sequence of a 99..then... :mad: :mad: :mad:
while I was out bush too....2 days crunching kaput!

[ohms18k]

bloody 28th sequence of a 99..then... while I was out bush too....2 days crunching kaput!

That's the one thing worst than a small gene is a big gene that crashes especially on sequence 30. What a waist of time.

FYI-Also had a race between 2 boxes win xp vs win 98, running 12aa genes over the weekend. the 98 box is faster but the xp box ran the most genes +15. the 98 box crashed Sunday. xp has not crashed the client once. As a matter of fact xp hasn't crash at all.....hmmmm.

Check out my mad gene count.

[Dave S]

Originally posted by ohms18k


That's the one thing worst than a small gene is a big gene that crashes especially on sequence 30. What a waist of time.

FYI-Also had a race between 2 boxes win xp vs win 98, running 12aa genes over the weekend. the 98 box is faster but the xp box ran the most genes +15. the 98 box crashed Sunday. xp has not crashed the client once. As a matter of fact xp hasn't crash at all.....hmmmm.

Check out my mad gene count.
man that is a lot of genome's 1444 WOW :eek:

[ohms18k]

man that is a lot of genome's 1444 WOW

Yeah i know........

Ok, no more small genes. That’s way out of norm :eek: and not much in units. Even though I can get more units out of small genes before the big gene crashes, the time it takes for me to find out it’s crashed, could be a couple of days of spinning it’s fans doing nothing. Going to go mid range and see how they stack up. I wish they fix the client and those stats at Stanford.

[Dave S]

Originally posted by ohms18k


Yeah i know........

Ok, no more small genes. That’s way out of norm :eek: and not much in units. Even though I can get more units out of small genes before the big gene crashes, the time it takes for me to find out it’s crashed, could be a couple of days of spinning it’s fans doing nothing. Going to go mid range and see how they stack up. I wish they fix the client and those stats at Stanford.
LOL
I know wot U mean about the client crashing I must of lost 200 to 400 units :eek:, that sucks – but hey I guess it’s all part of the fun:rolleyes:

[siggy]

Originally posted by Dave S

man that is a lot of genome's 1444 WOW :eek:

Its weird looking at your stats with your gene count so high, it looks out of wack......

[ohms18k]

Its weird looking at your stats with your gene count so high, it looks out of wack......

You know i kind like having stats out of whack. :D It shows TCG has all kinds of whacky stuff going on.

[Dave S]

Originally posted by ohms18k


You know i kind like having stats out of whack. :D It shows TCG has all kinds of whacky stuff going on.
LOL keeps the rest on their toes mate :D:D

[shubles]

i just got my first ever crash

87@25 :(

if u had my machine u'd know how badly this affects things.

say goodbye to about 2or 3 days crunching.

thanks dnar...
...somewhere in here you mentioned running "-clear"
that fixed the client, and it got me a 98aa.
this one better not f***ing crash!!!!!!!

[dnar]

Ooohh, the 98aa.... Yeh, they crash ALL the time!!!:D

[Dave S]

Originally posted by dnar
Ooohh, the 98aa.... Yeh, they crash ALL the time!!!:D
LOL she will come round after U m8;)

[shubles]

Originally posted by Dave S

LOL she will come round after U m8;)

THAT'S RIGHT!!!!!

i know where you live dnar!!!!!

*well sort of*

[siggy]

[QUOTE]Originally posted by shubles
[B]

THAT'S RIGHT!!!!!

i know where you live dnar!!!!!

Really Dnar I can only slap you around on the computer.....Shubles can do it in person.....

:p :p

[Dave S]

Originally posted by shubles


THAT'S RIGHT!!!!!

i know where you live dnar!!!!!

*well sort of*
LOL go get him girl :D

[dnar]

LOL! All these lovely women fighting over me, jeesh, I'm flatered.....



Dave_S - I notice your a "386" post. I still got one of them! LOL.



http://members.iinet.net.au/~dnar/images/icons/homer2.gif

[Dave S]

Originally posted by dnar
Dave_S - I notice your a "386" post. I still got one of them! LOL.
http://members.iinet.net.au/~dnar/images/icons/homer2.gif
WOT a 386 noway:confused: