Error occurred while writing report log: Could not find a part of the path ‘Z:\deadlineRepository\reports\jobs\bb\3\5273ca2d4a7a680b34882bb3.bz2’. (System.IO.DirectoryNotFoundException)
I am submitting a V-ray job from maya 2013. I have just upgraded to version 6 from 5 and haven’t seen an error like this before. The jobs folder listed is also empty.
The problem if you installed 6.0 on top of 5.2 is the fact that a 5.2 Slave will delete everything out of the logs folder during its housekeeping sweep. That includes the folders the Slaves require to write in. So! One of two things need to be done:
Prevent all the 5.2 Slaves from starting. Uninstalling may be good.
Make 6.0 use a completely different network path.
Then re-install the 6.0 Repository so the folders are all re-created.
I’ve asked Ryan to have an error triggered in the installers when doing a major upgrade into an old Repository. We’ve always required a fresh install between major versions, but this is the first one I can remember where it can do some damage to the new installation.
We’re having the same problem but DID uninstall 5.2 from each machine before installing 6.x. We did have to change where the repository was pointing, but it’s still erroring. All workstations, renders settings, render nodes, etc. are pointing to the correct Repository directory. We tried running the slaves locally without the service and it still is erroring. Any ideas as we can’t use it now?
So there are a few ways to verify what is happening here. If you have a rogue 5.2 instance, you will find a /slaves folder in your repository root. Inside there will be a folder with the hostname of the machine causing these issues. Another way to verify is if the “[repo]\reports\jobs” folder is empty, or does not have a series have 2 character folder names, such as 00, 0a, 0b, and so on. These are 2 pretty clear indications of a rogue 5.2 slave going through it’s cleanup process. If you take the hostname in the first example, it should show you which machine is gone bad.
Ah, yes… Sorry for the delay on getting back about this. I was so sure that we didn’t have any 5.2 installed that I didn’t check that folder. Well apparently we did have a rogue 5.2 installed. Of all the machines, it was on the machine I initially installed the 6.x database on. I must’ve just run the repository and not the 6.x installer.