Has anyone experienced or heard of the following problem:
- First we are running CVS 1.11 on Solaris 2.8
- Desktop checkin/checkout through WinCVS works great
- Unix checkin/checkout results in problems that show up as file corruption
on a machine reboot. We're trying to pinpoint the commands that might be
causing the problem, but here is what we've found so far:
-- we run some combination of "cvs export -r tag", "cvs checkout module",
plus "rm -R *" commands to clean out the working directories.
-- The next time our Solaris machine reboots, it reports file corruptions
that must be cleaned up with fsck.
-- The fsck activity results in a bunch of lost+found files being created
with the timestamp and permissions of the user who was doing the cvs work.
-- The contents of the lost+found files look like CVS controls information.
--- Example: A lot of files contain the strings "Entries", "Entries.log",
"Repository", "Root"
--- Example: Other files contain list of files in a given directory in the
repository
-- The CVS repository seemed unaffected by these problems -- it's come back
clean every time
-- Some of the file corruption occurs on the boot drive of the unix
machine. We need to clean it up first, then clean up the problems on the
filesystem that houses CVS and the CVS work directories.
Has anyone else experienced anything like this?
Any suggestions on paths that could help us resolve the situation?
Thanks,
-brian
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
