I believe I'm having some trouble with Qpopper 4.0.4. Odd things have been happening lately that I haven't been able to figure out. The box is an dual x86 box running RH 7.3. Qpopper was configured with
--prefix=/usr/local/ --sysconfdir=/etc/qpopper --with-pam=pop3 --with-popuid=pop --with-log-facility=LOG_LOCAL1 --enable-shy --enable-log-login --enable-servermode --enable-bulletins=/var/mail/bulls --enable-spool-dir=/var/mail/spool --enable-popuid=pop --enable-temp-drop-dir=/var/mail/poptemp --enable-fast-update when I migrated to the new server back in August. Over the past couple of weeks I received reports from one of the admins that he was receiving duplicate mail. I'm in the middle of configuring MIMEDefang w/ SpamAssassin in Sendmail as well and using procmail to forward a copy of all mail to a 2nd box for AV stress testing so I thought I might have goofed something up. I had also just changed the Sendmail config to the 2 daemon setup that 8.12 uses (whereas I'd been using the old single control process setup even with 8.12.6 for some time now). After some digging around I confirmed that the admin only had a single copy of the messages in his inbox. I pretty much figured the MUA (LookOut(tm)) was to blame. We nulled his spool and he configured LookOut to not leave mail on server and it went away. Of course it slowly but surely happened to all 4 of his accounts on the mail server. I suspect users are experiencing this too but haven't reported it. Tonight I noticed a large number of files in my mail spool directory named things like _1AD,25A39.oak. Well, here let me give you a sample: -r-------- 1 hoov smmsp 0 Nov 26 10:34 _06C,iK649.oak -r-------- 1 hoov smmsp 0 Nov 26 10:34 _06C.iK649.oak -r-------- 1 hoov smmsp 0 Nov 26 10:34 _06C%iK649.oak -r-------- 1 hoov smmsp 0 Nov 26 10:34 _06C+iK649.oak -r-------- 1 parsonsg smmsp 0 Nov 26 10:37 _0-D,DN649.oak -r-------- 1 parsonsg smmsp 0 Nov 26 10:37 _0-D.DN649.oak -r-------- 1 parsonsg smmsp 0 Nov 26 10:37 _0-D%DN649.oak -r-------- 1 parsonsg smmsp 0 Nov 26 10:37 _0-D+DN649.oak -r-------- 1 cjbach smmsp 0 Nov 20 16:40 _0DE,f9A39.oak -r-------- 1 cjbach smmsp 0 Nov 20 16:40 _0DE.f9A39.oak -r-------- 1 cjbach smmsp 0 Nov 20 16:40 _0DE%f9A39.oak -r-------- 1 cjbach smmsp 0 Nov 20 16:40 _0DE+f9A39.oak -r-------- 1 lukert smmsp 0 Nov 20 16:38 _0hD,Y7A39.oak -r-------- 1 lukert smmsp 0 Nov 20 16:38 _0hD.Y7A39.oak -r-------- 1 lukert smmsp 0 Nov 20 16:38 _0hD%Y7A39.oak -r-------- 1 lukert smmsp 0 Nov 20 16:38 _0hD+Y7A39.oak -r-------- 1 annl smmsp 0 Nov 29 12:41 _0IH,kT759.oak -r-------- 1 annl smmsp 0 Nov 29 12:41 _0IH.kT759.oak -r-------- 1 annl smmsp 0 Nov 29 12:41 _0IH%kT759.oak -r-------- 1 annl smmsp 0 Nov 29 12:41 _0IH+kT759.oak Sorry if it wraps badly. These files only date back to Nov 20. None of them have data in them. There are 2783 total. I thought it might be an openssl problem but I see that I'm running an ok version (OpenSSL 0.9.6g). Does anyone have any ideas what's up with this? I'm a bit stumped. I can't decide if these problems are related or not. Does the client keep track of where it's at in a spool or is that the POP daemons job? I can't remember. I'd like to better my qpopper setup sometime in the future. If anyone has any sugestions, I'd love to hear them. The spools are on disks local to the mail server but the home directories are NFS mounted. I thought about putting poptemp on a tmpfs volume. I need more ram for that though. Thanks Justin
