Re: How to handle mail for multiple (10-15) domains w/o localpart conflicts?
On Mon, 07 Apr 2003 12:12:46 +1000, Donovan Baarda wrote: [...] >> I don't want to move away from Exim unless there are serious (or even >> compelling) reasons to do so. > >This has to rate as a FAQ. Time to start documenting it in the wiki; > >http://wiki.debian.net/EmailConfiguration > >Includes documentation on configuring Debian for Maildir using procmail >with exim or postfix. Thank you for taking the time to provide this. Now that I've had a look at this I still don't see strong reasons NOT to use Exim any longer. I do see reasons to migrate to "maildir"-style mailboxes which I might do (altho as I mentioned performance probably won't EVER be an issue at all since the server will be running more or less "idle.") -- L I N U X .~. The Choice /V\ of a GNU /( )\ Generation ^^-^^
Re: How to handle mail for multiple (10-15) domains w/o localpart conflicts?
On Mon, 07 Apr 2003 10:13:43 +1000 (EST), Brad Lay wrote:
[...]
>I've just setup this exact same thing on my mail server. You don't need
>any debian packages for everything. Heres my configs.
Thanks for sharing this with us. As I already pointed out I would really like
a solution that can be realized without not-yet-Debianized software (even it's
unofficial packages that I have to use.)
[...]
> server_condition = "${if eq{${md5:$3}}{${substr_5:${lookup mysql{select
password_hash from popbox where local_part='${local_part:$2}' and
domain_name='${domain:$2}'}{$value{1}{0}}"
But how can I *easily* enter new users into MySQL? I'm not exactly a MySQL
expert... Is there a web interface for these kind of jobs? Maybe as part of
"vmail-sql?"
>Hope this helped. If you need more help shoot me an email offlist.
Well, I'm currently only collecting infos. I've still to rent my server (I'm
already running a private one since 5 years or so, but this time I'm providing
the service for "customers" (read: friends) in a somewhat larger scale (but
still it's small-scale for "real" ISPs ;-)
>[1] Debian's exim needs to be recompiled to use mysql (I can give you a
>url to the one I created).
That's no problem for me. I published an Exim4 package I created myself LONG
before there were even experimental official ones. But thanks for your offer,
and thanks for taking the time of writing this all up.
--
L I N U X .~.
The Choice /V\
of a GNU /( )\
Generation ^^-^^
Re: How to handle mail for multiple (10-15) domains w/o localpart conflicts?
On Sun, 06 Apr 2003 22:04:32 +0200, Markus Welsch wrote: >> Well, performance is not a problem for me. As I already mentioned I'm just >> hosting a dozen of domains with only a couple of (low-use) mailboxes >> altogether. > >Well that really doesn't make a difference, but nevermind ! Well, I think it does. Solutions that are perfect for my needs are probably just inacceptable for "real" ISPs (because of the lack of performance, comfort, billing facilities, etc.) >> Sure, but I like the option of quickly having a glance at them using "Mutt," >> in case something goes wrong etc. This way I can ssh in and remove offending >> mails, instead having to telnet to 110 and DELE it. > >Well you could also set up webmail if you don't like to telnet. That's right. I will be doing that, in fact, using OpenWebMail (which seems to run great UNLESS the home directories are on NFS, which won't be the case for me.) >What do you count to "offending mails" ? Well, HUGE messages which users can't (or don't want to) download, and which I can easily delete using Mutt on the mbox file. >> What exactly does the above do? I'm sorry but I don't quite get your point. > >Well I thought you meant just for some hosts e. g. the mailserver host >you would like to have local deliverage and for all others you would >like to use cyrus. Call me stupid but I still don't quite get you. Maybe you would like to explain me your point off-list? >I don't think not using mbox and rather using a better way is not a >mistake. And I wouldn't also limit the decision of which You're absolutely right, but as I said (because I'm quite familiar with Exim) I would like to stick with it unless there are serious (or even compelling) reasons to move away fromit. And "better performance" isn't a reason for me since the server will be 99% idle all the time. >POP3/IMAP-server to use just because a MUA can't handle the format of >the server ... Again you're absolutely right, and now that you have reminded me of webmail this point is no longer an issue for me. >> This option I will only consider if there is NO OTHER WAY. I've once started >> to do that and I have very soon regretted that I did so, so I don't want to >> repeat that mistake again. > >Those are just php scripts and a sql database creation script. You do >not need to compile anything, etc. All it checks for is if you have >webserver, php, etc installed. Ok, in that case I might use it if I don't find a better "environment" for my needs. Thanks, Ralf -- L I N U X .~. The Choice /V\ of a GNU /( )\ Generation ^^-^^
Re: dialin works only once or twice
Hi Jan, I had some problems with mgetty too, as i can remember ( he, its monday ) there are options to get more logging from pppd and mgetty. Setting more verbose logging will help you with finding the problem. Another thing you can try is to make a "cleanup" script in /etc/ppp/ip_down.d that makes sure the pppd is killed and the line is down and to run some checks ( its allso possible to add some custom call logging ;) ) If you can't find the problem feel free to send me a email grtnx, Robbert Helling Citeren Jan Harders <[EMAIL PROTECTED]>: > hi everybody, > > i have woody installed and am trying to build a dialup account for users > ... > no problem i couldn't have solved with a few docs ... > things are running just fine once, maybe twice, then there's no answer to > the incoming call any more. system uses an internal isdn-card (avm fritz > pci) and net_tty shows the call in the logs and rings on ttyI0, mgetty > seems > to take it > - > 04/06 18:02:43 yI0 mgetty: experimental test release 1.1.27-Oct21 > 04/06 18:02:43 yI0 check for lockfiles > 04/06 18:02:43 yI0 locking the line > 04/06 18:02:44 yI0 lowering DTR to reset Modem > 04/06 18:02:44 yI0 send: ATZ[0d] > 04/06 18:02:44 yI0 waiting for ``OK'' ** found ** > 04/06 18:02:44 yI0 send: AT&E0[0d] > 04/06 18:02:44 yI0 waiting for ``OK'' ** found ** > 04/06 18:02:44 yI0 send: AT&B512[0d] > 04/06 18:02:44 yI0 waiting for ``OK'' ** found ** > 04/06 18:02:45 yI0 waiting... > 04/06 18:03:18 # failed dev=ttyI0, pid=23235, got signal 15, exiting > > > but, as you can see, quits. couldn't find any additional information what > happens just before the signal 15 msg. the client (win 98) just tells me, > the remote computer didn't answer. tried the exact same config of mgetty, > isdn and ppp on another computer (with a teledat 100 isa isdn card) and it > works fine all the time. > any ideas? > if you need more information, please let me know. > > thanks in advance, > > jan harders > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact > [EMAIL PROTECTED] > > > __ http://www.wanadoo.nl/
[HELP] .htaccess problem
Hello List : i using .htaccess to restrict user to use web resource . However, i type correct username & password in box that the web browser pops up when i attempts to access resource in protected area. i got error message ● in /var/log/apache/error.log as following : [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin [Mon Apr 7 15:31:28 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin ● in /var/log/syslog : Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:07 backup slapd[17788]: conn=4 op=4 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:07 backup slapd[17788]: conn=4 op=4 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:09 backup slapd[17789]: conn=5 op=5 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" === ● My .htaccess as follow : AuthName "For Student to login" AuthLDAPUrl ldap://192.168.8.8/dc=ezplay,dc=tv?uid? AuthType Basic Order deny,allow Deny from all Allow from all require user test ● However, when i execute "ldapsearch" command , the result as follow: backup:/var/www/admin# ldapsearch -x -h 192.168.8.8 -b dc=ezplay,dc=tv uid=tester version: 2 # # filter: uid=tester # requesting: ALL # # search result search: 2 result: 0 Success # numResponses: 1 i am NO any idea to solve this problem, and i have been tried to find out answer in MaillistUnfortunately,i didnt got answer about my problem Please Help me. -- Trust & Unique ... axacheng <[EMAIL PROTECTED]>
Re: [HELP] .htaccess problem
Hi, may be it's very simple: [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): You typed in the username "tester", didn't you? ---^^ But in your .htaccess, you said that only the user "test" should be allowed to access the area: Order deny,allow Deny from all Allow from all require user test That's all, I think! Best regards, Andreas -- procommerz - Internet fuer Unternehmen http://www.procommerz.de | 033925-90710
Re: [HELP] .htaccess problem
Hello Andreas : thanks a lot but..Sorry...its my fault. my .htaccess as follow Order deny,allow Deny from all Allow from all require user tester so that, i still CAN NOT solve this problem @___@ i dont know what mean error message : [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): -- Trust & Unique ... axacheng <[EMAIL PROTECTED]>
Re: [HELP] .htaccess problem
Hi, you do actually use the objectclass attribute in your entry for uid=tester? (I think you are supposed to at least have objectclass=top) Regards, Cato Aune mandag 7. april 2003, 12:44, skrev axacheng: > Hello Andreas : > > thanks a lot > > but..Sorry...its my fault. > > my .htaccess as follow > > > > Order deny,allow > Deny from all > Allow from all > require user tester > > > so that, i still CAN NOT solve this problem @___@ > > > i dont know what mean error message : > > [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search > must return exactly 1 entry; found 0 entries for search > (&(objectclass=*)(uid=tester)): > > > > > > > -- > Trust & Unique ... > axacheng <[EMAIL PROTECTED]>
Re: [HELP] .htaccess problem
On Mon, 2003-04-07 at 12:44, axacheng wrote: > i dont know what mean error message : > > [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return > exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): > google says it's an ldap error. hav you got mod_auth_ldap loaded and not configured? tinus
Re: [HELP] .htaccess problem
> > [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must > > return exactly 1 entry; found 0 entries for search > > (&(objectclass=*)(uid=tester)): > google says it's an ldap error. > hav you got mod_auth_ldap loaded and not configured? sorry, didn't pay attention. tinus
Re: [HELP] .htaccess problem
Hmm, I'm not totally sure, because I haven't worked with this in some (Btime and I'm also not sure which LDAP authentication module you are (Busing, but my auth_ldap-1.6.0-4 configuration follows: (B (B (BAuthType Basic (B (BAuthName Login (B (BAuthLDAPURL (Bldap://localhost:389/ou=People,dc=yourdomain,dc=org?uid?sub (B (BAllowOverride AuthConfig (B (Brequire valid-user (B (B (BAlso, this configuration is included in httpd.conf for a (Bcontext. But you should be able to apply it to your case. (B (B (BHope that helps. (B (B (BEirik (B (B (BOn Monday, April 7, 2003, at 04:14 AM, axacheng wrote: (B (B (B (B$B!|(B My .htaccess as follow : (B (B (BAuthName "For Student to login" (B (BAuthLDAPUrl ldap://192.168.8.8/dc=ezplay,dc=tv?uid? (B (BAuthType Basic (B (B (B (B (BOrder deny,allow (B (BDeny from all (B (BAllow from all (B (Brequire user test (B (B (B (B (B (B
Privacy in virtual hosting environment
Hi, Since I'm currently setting up my first shared hosting environment in a few years I'm wondering how to adequately address privacy issues. I plan to provide python (with and without mod_python), perl (perhaps just CGI) as well as PHP support. Is there any way to prevent people from sourcing things above their document root? My main concern is db passwords stored in config files or scripts. I think with apache2 and the mpm that allows different users for different virtual hosts it should be possible but since apache2 isn't in woody I don't like the idea too much. My other thoughts are to run multiple instances of apache wth different uids, an alternative might be to run user mode linux or other virtual environment. Both of these options seem quite resource intensive though and multiple apaches would require an IP for every site (I think). What are you guys using? Thanks, -- Fraser Campbell <[EMAIL PROTECTED]> http://www.wehave.net/ Brampton, Ontario, Canada Debian GNU/Linux
Re: Privacy in virtual hosting environment
- Original Message - From: "Fraser Campbell" <[EMAIL PROTECTED]> To: Sent: Monday, April 07, 2003 7:36 AM Subject: Privacy in virtual hosting environment > Hi, > > Since I'm currently setting up my first shared hosting environment in a few > years I'm wondering how to adequately address privacy issues. Good, lots of 'em don't care. > I plan to provide python (with and without mod_python), perl (perhaps just > CGI) as well as PHP support. Is there any way to prevent people from > sourcing things above their document root? My main concern is db passwords > stored in config files or scripts. I tried using PHP's ini settings within each vhost to ensure people weren't access directories outside the directory for their vhost but that didn't work because PHP screws it all up and gets mixed up with some vhosts parameters bleeding into others and it generates errors in a seemingly random way. suexec is nice for cgi but still doesn't limit a scripts ability to read files outside the a certain directory. It only enforces permissions. So at first my solution to that was to restrict vhost directories by permissions and have a separate group for each vhost that has the apache user in it and then under that directory, everything is world readable. That way, only the user and apache have access to the primary directory and then standard permissions apply after that. So if suexec is running cgi as one user another can get outside their own vhost directory but no inside anyone elses. e.g. directory, owner, group, perms vhost1,user1, group1-apache, 750 vhost1/htdocs, user1, group1, 755 vhost2,user2, group2-apache, 750 vhost2/htdocs, user2, group2, 755 Well that breaks down as soon as apache is in more than the maximum allowed number of groups per user thus limiting you to that number of vhosts per apache instance. > I think with apache2 and the mpm that allows different users for different > virtual hosts it should be possible but since apache2 isn't in woody I don't > like the idea too much. that's finally what I've resorted to. Even though mpm is experimental I have to use it anyway because it really is the best option..USE THE SOURCE, LUKE..heheh > My other thoughts are to run multiple instances of apache wth different uids, > an alternative might be to run user mode linux or other virtual environment. > Both of these options seem quite resource intensive though and multiple > apaches would require an IP for every site (I think). been there done thatvery uglymost secure but very ugly. It is a pain to administer and it is pig on the box. Yes each one requires it's own IP address if you are going to have them all listen on port 80 anyway. > What are you guys using? > I know some that have switched to web servers like Roxen to help solve this issue and make things more secure. I don't know if PHP will run on stuff like that without running it as a CGI instead of having PHP built-in. I know of admins that have run separate instances of apache like you mention above. vec
Re: port blocking
On Saturday 05 April 2003 01:36 pm, [EMAIL PROTECTED] wrote: > I'm trying to lock down my server, which, for historical reasons *has* > to run the various nis services. No problem, I'll just block the ports > that ypfrx, yppasswdd, ypbind, etc bind to. However, it seems that they > choose a different port each time. As I don't want to switch to the > 'block everything, only open needed' methodology (too much overhead to > keep all my clients working), how do I force the various nis services to > use only certain specified ports? Looking at the man page, some of them > take a -p switch, but putting that into the /etc/init.d/nis file in the > --exec line 1) doesn't seem to work and 2) would be overwritten by the > next upgrade (iirc, the init scripts are not marked as config files). > Any ideas? I'm not exactly sure if locking the NIS ports is possible, but I can verify that init.d scripts are (almost?) always marked as conffiles. You do not need to worry about them being overwritten without being asked. That said, I would really recommend that you switch to the "block everything" firewall methodology, especially if you need legacy software like NIS around. You'll sleep easier. Also, about the -p switch: make sure you're passing "--" between the program name and the -p. start-stop-daemon needs this, to seperate its own arguments from those of the daemon it's starting. (This may not be your problem - like I said, I don't use NIS, those scripts might not call start-stop-daemon.) Hope this helps. - Keegan
[HELP] .htaccess problem.......thanks.
Very Thanks for all reply. :-) Now , i type correct username & password in box that the web browser pops up when i attempts to access resource in protected area . its still didn't work-- its made me a pretty BAD mess of this problem ...Help ... help.i wanna die... ● My LDAP tree as follow ： dc=ezplay,dc=tv | |>uid=tester | |>uid=axa.cheng | |__ou=td | | | |>uid=bigbrother | |>uid=bigcow | |__ou=md | | |>uid=freesec ● However, when i execute "ldapsearch" command , the result as follow: ldapsearch -W -x -h localhost -b "dc=ezplay,dc=tv" -D "uid=tester,dc=ezplay,dc=tv" '(&(objectclass=*)(uid=tester))' version: 2 # # filter: (&(objectclass=*)(uid=tester)) # requesting: ALL # tester, ezplay, tv dn: uid=tester,dc=ezplay,dc=tv uid: tester cn: tester objectClass: account objectClass: posixAccount objectClass: top objectClass: shadowAccount userPassword:: e2NyeXB0fSQxJDddf5VU0YU5SJEQvQjRd3kFVdkppNTFQsdsISzl5WS8= shadowLastChange: 12128 shadowMax: 9 shadowWarning: 7 loginShell: /bin/bash uidNumber: 1008 gidNumber: 100 homeDirectory: /home/tester # search result search: 2 result: 0 Success # numResponses: 2 # numEntries: 1 ^^ i got a entry now... :-) ● My .htaccess as follow : AuthName "For Student to login" AuthLDAPUrl ldap://192.168.8.8/dc=ezplay,dc=tv?uid? AuthType Basic Order deny,allow Deny from all Allow from all require user tester ● in /var/log/apache/error.log as following : [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin [Mon Apr 7 15:31:28 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin ● in /var/log/syslog : Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SEARCH RESULT tag=101 err=0 text= == Hello List : i using .htaccess to restrict user to use web resource . However, i type correct username & password in box that the web browser pops up when i attempts to access resource in protected area. i got error message ● in /var/log/apache/error.log as following : [Mon Apr 7 15:31:27 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin [Mon Apr 7 15:31:28 2003] [error] [client 192.168.10.254] Search must return exactly 1 entry; found 0 entries for search (&(objectclass=*)(uid=tester)): URI /admin ● in /var/log/syslog : Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:03 backup slapd[17788]: conn=3 op=4 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:05 backup slapd[17789]: conn=7 op=2 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:07 backup slapd[17788]: conn=4 op=4 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" Apr 7 15:50:07 backup slapd[17788]: conn=4 op=4 SEARCH RESULT tag=101 err=0 text= Apr 7 15:50:09 backup slapd[17789]: conn=5 op=5 SRCH base="dc=ezplay,dc=tv" scope=2 filter="(&(objectClass=*)(uid=tester))" === ● My .htaccess as follow : AuthName "For Student to login" AuthLDAPUrl ldap://192.168.8.8/dc=ezplay,dc=tv?uid? AuthType Basic Order deny,allow Deny from all Allow from all require user test ● However, when i execute "ldapsearch" command , the result as follow: backup:/var/www/admin# ldapsearch -x -h 192.168.8.8 -b dc=ezplay,dc=tv uid=tester version: 2 # # filter: uid=tester # requesting: ALL # # search result search: 2 result: 0 Success # numResponses: 1 i am NO any idea to solve this problem, and i have been tried to find out answer in MaillistUnfortunately,i didnt got answer about my problem Please Help me. -- Trust & Unique ... axacheng <[EMAIL PROTECTED]>
