Member of the EVE Tweet Fleet

Monday, June 21, 2010

More clues to the PI bug.

Once again I had some missing materials. Due to something I was checking prior to the losses registering, I think I have a clue as to what is happening. I also had an extraction cycle re-set it self in mid cycle. I.e. I had the planet doing a 23h extraction cycle on all extractors on the planet. I then did something last night and this morning when I went to do the extractors, they showed exactly 11h left to go when they should have been dead. The thing is that what I had done was setup a route for one of 3 materials to a high tech processor. Then before submitting it I hit cancel. Which should mean there would have been no effect. However, this morning, that specific material had 1 cycle of materials missing... After I had exported something else out of storage it was residing in...

This is building up to be a sequence of events to check to see if I can reproduce this event reliably. The good news is that at least after every loss I'm getting closer to being able to reproduce the bug. This is good. The fact it's tricky to reproduce explains how it got missed in development. So careful with that cancel button.

5 comments:

Unknown said...

Be sure to recheck your extrator routes also. Yesterday evening when I was restarting my extractors with their good 23h cycles (after doing 5h Cycles during the day sunday), exactly 1/4 of my extractor had their routes "Lost".

I don't know how it happened, but 1 out of 4 had their outgoing materials "unrounted" and I'm pretty sure i they were before that.

So I don't know if changing from a 23h cycle -> 5h cycle -> 23h cycle has a small chanche of deleting the outgoing route, but something wierd is going on

Unknown said...

You may have lost the route when going from 23h to 5h and not noticed it. It takes more m3 on your routes during the 5h runs.

Very easy, and not really a bug so much as no good "warning" that it couldn't continue to route.

IMarv

Unknown said...

Huh, that could very well be it... but wierd since the others way to overload a route gives a msg saying that this operation "can not be done because of heavy congestion"

Anonymous said...

I noticed the "time travel" effect as well this weekend. I log in to restart my extractors left on the 23 hour, and sure enough, they were still going for about 10 or so more hours. I didn't audit the amount as they are bulk extractors for export, so I don't know if the extractors just stopped along the way and restarted, or if they cycled with some weird time interval.

WH Industrialist said...

I've seen these bugs as well both the time travel one, we have all got the max time run on our extractors since we initially set them up, and have yet to kick them off again (this is from the night of the CC release)

also resources going missing I've confirmed it further down the production chain as I've been running 200 -> 60 batches of s2-s3 and they've been dropping out aswell. My current working theory has to do with submitting multiple processors off of the same storage facility at the same time as I believe there might be a conflict collecting resources from the storage.