Recovery from Journal

[Allen Lacy]

I am trying to recover some work that I did on a project this am by using journals (my desktop computer is crashing during Revit sessions as well as virus scans). I can get the journal to work on the desktop, but the computer crashes before completion.

I'm now trying to use a laptop with the same build of release 7 on it . I copied the journal file to the network, but I get an "unrecoverable error has occurred" message before the file is opened. What am I missing?

[sbrown]

While occassionaly you can get work back from a journal thats not their intent and if the journal runs through completion(the same place it crashed the first time) it will crash again. So I always open the journal file and delete the last command from it.

Typically journal files will work if all of the following criteria are met.

1. you have a copy of the file(rvt) in the exact same state it was in when the journal file started.(note the journal file starts the second you open revit and includes all commands including opening families and other files, so if you have opened files during the session, typically the journal will fail there because you most likely saved those so when it opens them they aren't in the same state.
2. it is not a worksets project(I've never seen a journal work on a worksets job. I guess its possible, but not likely)

3. no other files were openend during the session.

But there are many commands that just stop a journal, I don't know why.

[FK]

Journals will work on everything (we use them for debuggung after all), but trying to bring the central file back to a known condition is probably more work than redoing your local changes.

[sbrown]

When you say they will work on everything what do you mean. They very rarely work all the way thru. Typically a crash occurs after your 7th hour of work that day and the backups are far from the original state. I have asked before and will ask again, it would be great if the journals stopped and started over with each save, not with each session of revit. This would pretty much guarantee them working 9 times out of 10. They would also be much smaller. Some of my journal files will get to be 5 - 10 mb.

[Allen Lacy]

I'm pretty sure that my problem on the desktop computer was a hardware issue. I could get the journal to work, it's just that my computer is dying. Just wondered if switching to a different computer with the journal file was the reason it wouldn't work.

[FK]

I mean journals will replay more or less everything Revit does, including workest operations, given identical initial conditions, not that they'll be a reliable backup.

[LRaiz]

When you say they will work on everything what do you mean. They very rarely work all the way thru. Typically a crash occurs after your 7th hour of work that day and the backups are far from the original state. I have asked before and will ask again, it would be great if the journals stopped and started over with each save, not with each session of revit. This would pretty much guarantee them working 9 times out of 10. They would also be much smaller. Some of my journal files will get to be 5 - 10 mb.
If you know what you are doing you might be able to cut the the top portion of your journal up to the last successful save and replace it with a segment that retrieves last saved version and opens appropriate views/sheets. Journals will usualy work if initial conditions did not change. Saving a file in the middle changes initial conditions, so to workaround the problem one can try to replace an outdated portion of the journal.

[irwin]

I mean journals will replay more or less everything Revit does, including workest operations, given identical initial conditions, not that they'll be a reliable backup.
If, during the original session you had several interactions with the central file (e.g., two saves to central), and another user changed the central file between those operations (e.g., he saved to central), then the journal won't run regardless of which version of central you start from. In order for it to run, another user would have to change the central file in a precise way at a particular time interval during journal replay.

[sbrown]

So do you think its reasonable/possible to ask for a journal file that is meant to be used for recovery of lost work, vs. more of a development tool?

[FK]

Irwin: that's roughly what I meant by "too much work to bring central file to known condition". Not really known dynamically changing condition makes it worse indeed. ;-)

sbrown: journal files have always been a debugging tool. Saving the data is a saner way to assure recovery.