Thu Mar 15 21:08:08 UTC 2007
On Thu, 2007-03-15 at 12:47 -0700, Vernon Schryver wrote: > > From: Sven Willenberger > > > I will probably go ahead and do a fresh install of the dccd > > from scratch (versus updatedcc) just to see if that helps. > > updatedcc is just a shell script that does > unsetenv stuff > setenv other-stuff > ./configure conf-stuff > make install > /var/dcc/libexec/rcDCC start > > where the "stuff", "other-stuff", and "conf-stuff" are values captured > by ./configure the last time ./configure ran and built updatedcc. > > `updatedcc -x` may spew too much to be decoded, but > `/bin/sh -x updatedcc` should be clear. > > Or use an editor to look at the commands after this comment near the start > of updatedcc: > > # The following environment variables are set by this script to their values > > and the last 10% where updatedcc runs ./configure and make. > > Something must have happened during the patch installs on the solaris dcc-cl-a. I force-reinstalled the applicable patches, rebooted, deleted all the dcc* directories/files in /var/dcc/build and ran ./updatedcc -e "CC=gcc" (for some reason the link between cc and gcc on the system went byebye during the patch installs) Except for a "close" error by the tar command, it built cleanly, did a rebuild of the hash database and all are communicating again. All three are currently running with "-f" in the command line; I will change this back to the default in the next day or after making sure the OS and filesystems are stable.
More information about the DCC