Roxanne Sandesara wrote: > > On May 11, 2008, at 11:53 AM, Eric Shubert wrote: >> There are some other interesting options as well: >> -C : path for default config files >> Aha! I'm guessing that if you were to add "-C /etc/mail/spamassassin" >> to the >> exec command, that would solve your problem without having to use the >> /var/tmp/ directory. It doesn't explain though why your spamd is >> defaulting >> to /var/tmp first while other toasters are not. It does seem to be a >> simpler >> work-around though. > > OK. Made a copy of the run file for spamd as run.bak. Modified run to > add -C /etc/mail/spamassassin. Saved that. Made a copy of it as run.new > so that I could potentially switch back and forth, and have referents > before and after the change. > > Stopped qmail. Deleted the /var/tmp/spamassassin-toaster-root directory > and contents. (Since we want to see how this will work with the change > in place, we don't want there to be any chance it is reading the > 'stopgap' copy of config files I placed there to make things run > yesterday.) > > Started qmail. Checked our spamd log. The errors I was getting before > are no longer present. It is scanning successfully. The directory > /var/tmp/spamassassin-toaster-root is not being re-created. > > Obviously, this doesn't yet answer the question of how or why the system > was behaving as it was - and why uninstalling and re-installing the > packages didn't correct that issue - but adding the -C option to the run > file has at least corrected the problem so that the system is running > 'as it should' from its default configuration files without errors.
Thanks for the writeup Roxie. This gives a clear picture of what you did and the results. >> Is it possible that for some reason the vpopmail user doesn't have >> permissions to read /etc/mail/spamassassin and/or it's contents? >> # su - vpopmail -s /bin/sh >> # ls -l /etc/mail/spamassassin >> All of the files in this directory should be rw-r--r-- > > # su - vpopmail -s /bin/sh > -sh-3.00$ ls -l /etc/mail/spamassassin > total 52 > -rw-r--r-- 1 root root 1299 May 10 16:34 init.pre > -rw-r--r-- 1 root root 1299 May 8 19:02 init.pre.rpmnew > -rw-r--r-- 1 root root 399 May 10 16:34 local.cf > -rw-r--r-- 1 root root 399 May 8 19:02 local.cf.rpmnew > -rw-r--r-- 1 root root 1703 Jan 8 16:41 local.cf.rpmsave > -rw-r--r-- 1 root root 2397 May 10 16:34 v310.pre > -rw-r--r-- 1 root root 2397 May 8 19:02 v310.pre.rpmnew > -rw-r--r-- 1 root root 2443 May 7 15:27 v310.pre.rpmsave > -rw-r--r-- 1 root root 1195 May 10 16:34 v312.pre > -rw-r--r-- 1 root root 1195 May 8 19:02 v312.pre.rpmnew > -rw-r--r-- 1 root root 2416 May 10 16:34 v320.pre > -rw-r--r-- 1 root root 2416 May 8 19:02 v320.pre.rpmnew > I realized after asking this that permissions were not a problem, as earlier to had run spamassassin successfully after a su - vpopmail. It's interesting (to me) to see both .rpmnew and .rpmsave files. This indicates that the (noreplace) option was added to the spec file some time in between. > > So. That being the case of our results, it seems clear to me that there > is something 'hanging around' that is influencing both Spamassassin and > spamd to go looking for that directory. Initiating spamd with the -C > option is correcting for this. Which I am fine with, frankly, as I can > read man pages well enough to continue replicating that correction in > the future as needed. But it doesn't quite satiate my curiosity. Any > thoughts as to where I should look to run down where this might be > coming from? Or should I feed my curiosity some chocolate and tell it to > be happy as-is? :) > I found the bug: https://issues.apache.org/SpamAssassin/show_bug.cgi?id=5904 I've applied the work around to the spec file and it appears to be working. EE will put it through his tests and if it passes will post it to the site. Isn't life great on the bleeding edge? ;) -- -Eric 'shubes' --------------------------------------------------------------------- QmailToaster hosted by: VR Hosted <http://www.vr.org> --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
