Having seen the transitioner backlog was 40 years, I wondered how something could be outstanding from the year 1969.

So I looked up and found the following detail:-

---------------------
Transitioner backlog (hours) - this shows the number of hours behind the \"transitioner\" is, the transisioner is a program which keeps track of the changing states of workunits, keeping the database updated with the current state, what this actually shows is the age (from the current time) of the *oldest* workunit waiting to be changed to another state, it should be 0 when everything is good
-------------------------

Thinking that I could not understand that, I double checked the Project status again, and lo and behold, 40 years work seemed to have been done in 5 minutes --- it was back to zero!!!!

What happened?

Keith

Some part of the server software was behind in it\'s work, and then caught up.
Possibly something to do with uploader1.atm being turned off while data was being moved, and then turned back on and having a different amount of data registered, which the transitioner software had to catch up with.
It\'s currently at 347,397 hours.

Do not adjust your mind while we alter reality.

Some part of the server software was behind in it\'s work, and then caught up.
Possibly something to do with uploader1.atm being turned off while data was being moved, and then turned back on and having a different amount of data registered, which the transitioner software had to catch up with.
It\'s currently at 347,397 hours.

Do not adjust your mind while we alter reality.

Les

That is very, very strange, because it was that very same figure (347,397 hours = 40 years) when I looked before at about 20:15 UTC, but it was zero at 20:25 UTC on another look.
And it must have been back at 21:05 UTC to the 40 year figure when you looked!!!

Nothing could be outstanding from a time before the project even started.
My bet is that some date figure was set to zero during the data move, which could equate to 1st January 1970 (which is the date returned on computers with dead back up batteries).

Keith

Carl\'s working on the database (again!) at this moment so that could account for strange happenings.

---

Cpdn news

Keith
You\'re looking at it the wrong way.
It\'s not time-since-the-project-started. It\'s the total number of hours that will be required to get through the work load, starting from ..... NOW.
Except that it doesn\'t take that long. I don\'t know the details of how the figure is arrived at, but I\'ve seen a similar number in the past, and the work (whatever THAT is), had been dealt with a few hours later.

The server is just having an LSD flashback from Woodstock.

---

Keith
You\'re looking at it the wrong way.
It\'s not time-since-the-project-started. It\'s the total number of hours that will be required to get through the work load, starting from ..... NOW.
Except that it doesn\'t take that long. I don\'t know the details of how the figure is arrived at, but I\'ve seen a similar number in the past, and the work (whatever THAT is), had been dealt with a few hours later.

Les

Thanks for that explanation.
It is contrary to what is shown when I did a search:--
---------------------
Transitioner backlog (hours) - this shows the number of hours behind the \"transitioner\" is, the transisioner is a program which keeps track of the changing states of workunits, keeping the database updated with the current state, what this actually shows is the age (from the current time) of the *oldest* workunit waiting to be changed to another state, it should be 0 when everything is good
----------------------

So it seems that that needs to be corrected.
Maybe \"largest\" or \"longest\" would be more appropriate than \"oldest\" there.
No doubt you can see how I was misled.

Keith

The Seti server status page explains at the bottom what the various server programs do, though not all of theirs are used at CPDN.

Our transitioner backlog is now down to zero while the number of \'results in progress\' is down by nearly 200,000 since last night.... So that\'s always a good web page to look at when you need some thrills and spills.

---

Cpdn news

Just found this thread on google after noticing the transitioner backlog figure for the first time. Went from over 300,000 hours to 0 while I looked this up. Reminds me of some software installations where the time left to completion can be wildly inaccurate at some points of the operation.

Dave