Yesterday, I spent most of the day tracking down and fixing issues with the new cell creation scripts in the Debian package.
Lars Schimmer <[EMAIL PROTECTED]> writes: > I just setup a new cell on debian using the scripts in the package. > For all: it was NOT as easy as just running the scripts. > First: in the afs-newcell I added a line to add the IP to the > server/CellServDB. Without that entry, bos couldn't find the local > server. This is fixed in Subversion and the fix will be in the next upload. > After that Script in the server dir of preferences the CellServDB and > ThisCell was on localcell, maybe Ok, but I changed it to the real data. Your OpenAFS client has to be pointing to the local cell in order for afs-rootcell to work properly. This is documented in the afs-rootcell prerequisites. Why did you have to change it to point to some other cell? I'm not sure that I understand the problem that you were running into. > And for debian specific: I told the openafs-client not to start while > debconf installed it. That has to be changed after running that script, > or the afs-rootvol didn't run. Er, why? afs-newcell runs force-start after it finishes so that the OpenAFS client is running regardless of your debconf preferences. > OK, I managed to entry a user without the /admin tab from krb5 as > afs-admin, so I had problems getting admin token... I don't understand this sentence. Could you rephrase? I wrote a fairly complete document to accompany the configuration transcript that will also be in the next release of the package. I think that will help by giving more context for what's going on and discussing some of the possible alternative paths. -- Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/> _______________________________________________ OpenAFS-info mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-info
