Glenn Lagasse wrote: > * Danek Duvall ([email protected]) wrote: > > David Sechrest wrote: > > > > > This is the only thing under /tmp/m/var - > > > > > > % ls -l /tmp/m/var/user/root/tpm/userps/user.data > > > -rw------- 1 root root 0 Oct 7 14:13 > > > /tmp/m/var/user/root/tpm/userps/user.data > > > > > > Not sure how this came to be or what it means, or how best to clean it up. > > > > The security folks will know best, but it's completely useless where it is. > > You can move it to the dataset mounted at /var, but if there's something > > already there, I'd leave it alone and just blow this copy away (or save it > > aside and then blow it away). > > > > Point is, /tmp/m/var should be an empty directory when you're done. > > > > Now, the question of how this got to be there and not prevent /var from > > being mounted, is something I can't answer. > > Well, if recent data holds true, it's likely yet another SMF service > that starts up before filesystem-minimal is finished mounting the > datasets (in particular the var dataset) it needs to. The cluster folks > just ran into this and had to add a dependancy on their SMF startup > script to make things work in a world where /var is a separate dataset.
It might be trousers (tcsd), but that's not the path where TPM stores user data, at least not on my machine. On the other hand, the user tpm data on my machine is pretty old, so it may have changed. At any rate, the tcsd service has only a dependency on milestone/network, so if it does happen to write stuff to the filesystem, it's missing a dependency on filesystem/minimal. Copying Wyllys on this. Danek _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
