mail disaster recovery
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains. I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately. I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files? Any other pointers or "gotcha"s to this whole process? Dick
I'm not sure I completely understand. You have a few domains that are hosted on a server at your home that has DSL for the connection and it's down? If that is the situation then once you change the MX records to point to another server just make sure that the server you are pointing to will accept mail for that virtual domain. Also make sure that you pull your MX records from the server that is down so when it comes back up there are no conflicts. I don't know sendmail.cf that well, I use Qmail here that hosts over 50 different virtual domains with no problems. If that doesn't answer your question then let me know. At 02:56 PM 10/1/2003 -0400, you wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access type make in the /etc/mail/ directory and restart sendmail. you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain. when it is buffering the mail, it should show up in /var/spool/mail if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can. On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
-- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø How can you work when the system's so crowded?
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line: my-host MX 99 temp.server.domain. <--dot at end and restarted that named. Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed? What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address. Dick At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
do a: dig -t MX domain.com should show you if the DNS is reporting the backup mail sever correctly. On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
-- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø People are always available for work in the past tense.
; <<>> DiG 9.2.1 <<>> -t MX mydomain.net ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3412 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;mydomain.net. IN MX ;; AUTHORITY SECTION: mydomain.net. 2924 IN SOA mydomain.net. rlg.bach.mydomain. net. 2003100202 10800 3600 604800 3600 ;; Query time: 7 msec ;; SERVER: 67.100.202.xxx#53(67.100.202.xxx) ;; WHEN: Thu Oct 2 13:47:15 2003 ;; MSG SIZE rcvd: 75 I'd already gotten that output but didn't see any MX output, just the question. Your interpretation? Further suggestions? I'm losing mail at this point, but would like to stop the hemorraging. The SOA serial# indicates it is the most recent version of the modified (secondary) DNS Dick At 01:37 PM 10/2/2003 -0400, you wrote:
do a: dig -t MX domain.com
should show you if the DNS is reporting the backup mail sever correctly.
On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
you got 0 answer for your MX. your DNS is misconfigured. dig -t mx domain.com should in your case display at least 2 answers. your main server, and your backup server for example: dig -t mx wlug.org ; <<>> DiG 9.2.1 <<>> -t mx wlug.org ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63144 ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 3, ADDITIONAL: 0 ;; QUESTION SECTION: ;wlug.org. IN MX ;; ANSWER SECTION: wlug.org. 259200 IN MX 20 mx.WPI.EDU. wlug.org. 259200 IN MX 10 smtp.WPI.EDU. ;; AUTHORITY SECTION: wlug.org. 259200 IN NS ns1.wlug.org. wlug.org. 259200 IN NS alum.WPI.EDU. wlug.org. 259200 IN NS dns1.dovetailinternet.net. ;; Query time: 110 msec ;; SERVER: 192.168.2.48#53(192.168.2.48) ;; WHEN: Thu Oct 2 14:02:48 2003 ;; MSG SIZE rcvd: 149 You might need to set your secondary DNS to be a master for your domain. remember to restart the named. On Thu, 2 Oct 2003, Richard Goodman wrote:
; <<>> DiG 9.2.1 <<>> -t MX mydomain.net ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3412 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
;; QUESTION SECTION: ;mydomain.net. IN MX
;; AUTHORITY SECTION: mydomain.net. 2924 IN SOA mydomain.net. rlg.bach.mydomain. net. 2003100202 10800 3600 604800 3600
;; Query time: 7 msec ;; SERVER: 67.100.202.xxx#53(67.100.202.xxx) ;; WHEN: Thu Oct 2 13:47:15 2003 ;; MSG SIZE rcvd: 75
I'd already gotten that output but didn't see any MX output, just the question. Your interpretation? Further suggestions? I'm losing mail at this point, but would like to stop the hemorraging. The SOA serial# indicates it is the most recent version of the modified (secondary) DNS
Dick
At 01:37 PM 10/2/2003 -0400, you wrote:
do a: dig -t MX domain.com
should show you if the DNS is reporting the backup mail sever correctly.
On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
-- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø People are always available for work in the past tense.
After 1-1/2 hours studying Oreilly/Sendmail & Oreilly/DNS&Bind, I stumbled upon the correct format for my MX records ... turns out they were all wrong on a dozen domains on four servers. Now I have a new problem ... with the MX records working, the backup server is starting to receive mail, but rejecting it all for "Unknown user" ... I thought the mail was to be queued for redelivery, not attempted to deliver to mailboxes on the backup server. What have I done wrong NOW? Dick At 02:04 PM 10/2/2003 -0400, you wrote:
you got 0 answer for your MX. your DNS is misconfigured.
dig -t mx domain.com
should in your case display at least 2 answers. your main server, and your backup server
for example: dig -t mx wlug.org
; <<>> DiG 9.2.1 <<>> -t mx wlug.org ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63144 ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 3, ADDITIONAL: 0
;; QUESTION SECTION: ;wlug.org. IN MX
;; ANSWER SECTION: wlug.org. 259200 IN MX 20 mx.WPI.EDU. wlug.org. 259200 IN MX 10 smtp.WPI.EDU.
;; AUTHORITY SECTION: wlug.org. 259200 IN NS ns1.wlug.org. wlug.org. 259200 IN NS alum.WPI.EDU. wlug.org. 259200 IN NS dns1.dovetailinternet.net.
;; Query time: 110 msec ;; SERVER: 192.168.2.48#53(192.168.2.48) ;; WHEN: Thu Oct 2 14:02:48 2003 ;; MSG SIZE rcvd: 149
You might need to set your secondary DNS to be a master for your domain. remember to restart the named.
On Thu, 2 Oct 2003, Richard Goodman wrote:
; <<>> DiG 9.2.1 <<>> -t MX mydomain.net ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3412 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
;; QUESTION SECTION: ;mydomain.net. IN MX
;; AUTHORITY SECTION: mydomain.net. 2924 IN SOA mydomain.net. rlg.bach.mydomain. net. 2003100202 10800 3600 604800 3600
;; Query time: 7 msec ;; SERVER: 67.100.202.xxx#53(67.100.202.xxx) ;; WHEN: Thu Oct 2 13:47:15 2003 ;; MSG SIZE rcvd: 75
I'd already gotten that output but didn't see any MX output, just the question. Your interpretation? Further suggestions? I'm losing mail at this point, but would like to stop the hemorraging. The SOA serial# indicates it is the most recent version of the modified (secondary) DNS
Dick
At 01:37 PM 10/2/2003 -0400, you wrote:
do a: dig -t MX domain.com
should show you if the DNS is reporting the backup mail sever correctly.
On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site. That RH 7.3 machine runs mail/web/primary DNS for those domains.
I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
Any other pointers or "gotcha"s to this whole process?
Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
--
¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø
People are always available for work in the past tense.
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
sorry, the local-host-names file should be empty.. only the access file should have RELAY. it is trying to recive mail on your host, but the accounts dont exist on that box. On Thu, 2 Oct 2003, Richard Goodman wrote:
After 1-1/2 hours studying Oreilly/Sendmail & Oreilly/DNS&Bind, I stumbled upon the correct format for my MX records ... turns out they were all wrong on a dozen domains on four servers.
Now I have a new problem ... with the MX records working, the backup server is starting to receive mail, but rejecting it all for "Unknown user" ... I thought the mail was to be queued for redelivery, not attempted to deliver to mailboxes on the backup server. What have I done wrong NOW?
Dick
At 02:04 PM 10/2/2003 -0400, you wrote:
you got 0 answer for your MX. your DNS is misconfigured.
dig -t mx domain.com
should in your case display at least 2 answers. your main server, and your backup server
for example: dig -t mx wlug.org
; <<>> DiG 9.2.1 <<>> -t mx wlug.org ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63144 ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 3, ADDITIONAL: 0
;; QUESTION SECTION: ;wlug.org. IN MX
;; ANSWER SECTION: wlug.org. 259200 IN MX 20 mx.WPI.EDU. wlug.org. 259200 IN MX 10 smtp.WPI.EDU.
;; AUTHORITY SECTION: wlug.org. 259200 IN NS ns1.wlug.org. wlug.org. 259200 IN NS alum.WPI.EDU. wlug.org. 259200 IN NS dns1.dovetailinternet.net.
;; Query time: 110 msec ;; SERVER: 192.168.2.48#53(192.168.2.48) ;; WHEN: Thu Oct 2 14:02:48 2003 ;; MSG SIZE rcvd: 149
You might need to set your secondary DNS to be a master for your domain. remember to restart the named.
On Thu, 2 Oct 2003, Richard Goodman wrote:
; <<>> DiG 9.2.1 <<>> -t MX mydomain.net ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3412 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
;; QUESTION SECTION: ;mydomain.net. IN MX
;; AUTHORITY SECTION: mydomain.net. 2924 IN SOA mydomain.net. rlg.bach.mydomain. net. 2003100202 10800 3600 604800 3600
;; Query time: 7 msec ;; SERVER: 67.100.202.xxx#53(67.100.202.xxx) ;; WHEN: Thu Oct 2 13:47:15 2003 ;; MSG SIZE rcvd: 75
I'd already gotten that output but didn't see any MX output, just the question. Your interpretation? Further suggestions? I'm losing mail at this point, but would like to stop the hemorraging. The SOA serial# indicates it is the most recent version of the modified (secondary) DNS
Dick
At 01:37 PM 10/2/2003 -0400, you wrote:
do a: dig -t MX domain.com
should show you if the DNS is reporting the backup mail sever correctly.
On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
on the backup RH mail box: add you domain to the backup /etc/mail/local-host-names and set it to relay in /etc/mail/access
type make in the /etc/mail/ directory and restart sendmail.
you can try testing it. send mail to your domian. also, could telnet to mailserver.com 25 and put SMTP comands to verify it is accepting mail for that domain.
when it is buffering the mail, it should show up in /var/spool/mail
if your home DSL is going to be down for some time, you might just want to bring the box into the office, and point the DNS directly at it, if you can.
On Wed, 1 Oct 2003, Richard Goodman wrote:
> My home DSL has been down for over 48 hours and it appears that it will be > down for several more days. (I spare you the gory details). I am concerned > about losing mail coming into various domains at that site. > That RH 7.3 machine runs mail/web/primary DNS for those domains. > > I have modified the secondary DNS records for those domains (on a machine > at my office - also on DSL) with an higher cost MX record pointing to a > machine at the office also running RH 7.3 and sendmail, which I assume will > pick up and queue the mail. The DNS TTLs are reasonably short, fortunately. > > I've done this following O'Reilly/Sendmail book 2nd edition, but its not > always crystal clear if you don't already understand it! Is there anything > else I have to do on the backup machine (pointed to by the new MX records), > particularly is there anything I have to do to its sendmail config files? > > Any other pointers or "gotcha"s to this whole process? > > Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
--
¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø
People are always available for work in the past tense.
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
-- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Beggar to well-dressed businessman: "Could you spare $20.95 for a fifth of Chivas?"
Many thanks, looks like the mail server backup is operational. The first two pieces of incoming mail after the fix were spam to invalid addresses at the domain, but they were properly queued by the backup server in mqueue. Anything special about getting the mail out when my home DSL (and server) come back up in a few days. I was thinking of possibly moving the queued mail to a temporary directory a few times a day to cut down on needless retries before the DSL is restored. Any thoughts? Dick At 03:57 PM 10/2/2003 -0400, you wrote:
sorry,
the local-host-names file should be empty.. only the access file should have RELAY.
it is trying to recive mail on your host, but the accounts dont exist on that box.
On Thu, 2 Oct 2003, Richard Goodman wrote:
After 1-1/2 hours studying Oreilly/Sendmail & Oreilly/DNS&Bind, I stumbled upon the correct format for my MX records ... turns out they were all wrong on a dozen domains on four servers.
Now I have a new problem ... with the MX records working, the backup server is starting to receive mail, but rejecting it all for "Unknown user" ... I thought the mail was to be queued for redelivery, not attempted to deliver to mailboxes on the backup server. What have I done wrong NOW?
Dick
At 02:04 PM 10/2/2003 -0400, you wrote:
you got 0 answer for your MX. your DNS is misconfigured.
dig -t mx domain.com
should in your case display at least 2 answers. your main server, and your backup server
for example: dig -t mx wlug.org
; <<>> DiG 9.2.1 <<>> -t mx wlug.org ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 63144 ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 3, ADDITIONAL: 0
;; QUESTION SECTION: ;wlug.org. IN MX
;; ANSWER SECTION: wlug.org. 259200 IN MX 20 mx.WPI.EDU. wlug.org. 259200 IN MX 10 smtp.WPI.EDU.
;; AUTHORITY SECTION: wlug.org. 259200 IN NS ns1.wlug.org. wlug.org. 259200 IN NS alum.WPI.EDU. wlug.org. 259200 IN NS dns1.dovetailinternet.net.
;; Query time: 110 msec ;; SERVER: 192.168.2.48#53(192.168.2.48) ;; WHEN: Thu Oct 2 14:02:48 2003 ;; MSG SIZE rcvd: 149
You might need to set your secondary DNS to be a master for your domain. remember to restart the named.
On Thu, 2 Oct 2003, Richard Goodman wrote:
; <<>> DiG 9.2.1 <<>> -t MX mydomain.net ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3412 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
;; QUESTION SECTION: ;mydomain.net. IN MX
;; AUTHORITY SECTION: mydomain.net. 2924 IN SOA mydomain.net. rlg.bach.mydomain. net. 2003100202 10800 3600 604800 3600
;; Query time: 7 msec ;; SERVER: 67.100.202.xxx#53(67.100.202.xxx) ;; WHEN: Thu Oct 2 13:47:15 2003 ;; MSG SIZE rcvd: 75
I'd already gotten that output but didn't see any MX output, just the question. Your interpretation? Further suggestions? I'm losing mail at this point, but would like to stop the hemorraging. The SOA serial# indicates it is the most recent version of the modified (secondary) DNS
Dick
At 01:37 PM 10/2/2003 -0400, you wrote:
do a: dig -t MX domain.com
should show you if the DNS is reporting the backup mail sever correctly.
On Thu, 2 Oct 2003, Richard Goodman wrote:
Karl, That was exactly what I was looking for. Did the local-host-names changes; I already had the RELAY lines in access on the backup machine. Since the primary DNS is on the down IP address, I modified the secondary (slave) DNS which is one of my office boxes by adding a second MX line:
my-host MX 99 temp.server.domain. <--dot at end and restarted that named.
Its been quite a while and I'm not seeing any mail queueing, and I know there's mail that should be coming in on a regular basis. What else might I have missed?
What can I use (nslookup, dig ? with what parameters) to see whether the MX lines are right? When I send a test email from an off-site machine, the sending sendmail log says "Connection timed out" and shows the down IP address.
Dick
At 03:52 PM 10/1/2003 -0400, you wrote:
>on the backup RH mail box: >add you domain to the backup /etc/mail/local-host-names and set it >to relay in /etc/mail/access > >type make in the /etc/mail/ directory and restart sendmail. > >you can try testing it. send mail to your domian. also, could telnet >to mailserver.com 25 and put SMTP comands to verify it is accepting >mail for that domain. > >when it is buffering the mail, it should show up in /var/spool/mail > >if your home DSL is going to be down for some time, you might just >want to bring the box into the office, and point the DNS directly at >it, if you can. > > >On Wed, 1 Oct 2003, Richard Goodman wrote: > > > My home DSL has been down for over 48 hours and it appears that it will be > > down for several more days. (I spare you the gory details). I am concerned > > about losing mail coming into various domains at that site. > > That RH 7.3 machine runs mail/web/primary DNS for those domains. > > > > I have modified the secondary DNS records for those domains (on a machine > > at my office - also on DSL) with an higher cost MX record pointing to a > > machine at the office also running RH 7.3 and sendmail, which I assume > will > > pick up and queue the mail. The DNS TTLs are reasonably short, fortunately. > > > > I've done this following O'Reilly/Sendmail book 2nd edition, but its not > > always crystal clear if you don't already understand it! Is there > anything > > else I have to do on the backup machine (pointed to by the new MX > records), > > particularly is there anything I have to do to its sendmail config files? > > > > Any other pointers or "gotcha"s to this whole process? > > > > Dick
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
--
¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø
People are always available for work in the past tense.
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
--
¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl@hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 http://karl.hiramoto.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø
Beggar to well-dressed businessman: "Could you spare $20.95 for a fifth of Chivas?"
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
"Richard Goodman" <r.goodman@11harvard.com> wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site.
I was in a similar situation not too long ago when prolonged outages, address changes and ISP port blocking became too much to deal with. Losing my mail was fine, but the family went nuts when I was on the road and we had an outage. I wound up moving my SMTP to a $2/mo. hosting plan that will queue mail for me until I poll using POP via fetchmail. I've permanently pointed my MX record there, as their reliability is far better than mine (residential cable). I know this isn't as attractive as doing it yourself, and it might not be an option if you have a lot of domains to deal with. But it might be a quick fix if that outage is going to go on for a long time.
[...] I have modified the secondary DNS records for those domains (on a machine at my office - also on DSL) with an higher cost MX record pointing to a machine at the office also running RH 7.3 and sendmail, which I assume will pick up and queue the mail. The DNS TTLs are reasonably short, fortunately.
If the TTLs are short, hopefully senders will queue it until your backup is in place.
I've done this following O'Reilly/Sendmail book 2nd edition, but its not always crystal clear if you don't already understand it! Is there anything else I have to do on the backup machine (pointed to by the new MX records), particularly is there anything I have to do to its sendmail config files?
As Mike mentioned, there is configuration required on the backup server to accept for your domains. Nice to have an employer that doesn't mind that though!
Any other pointers or "gotcha"s to this whole process?
I was having fun doing it all via postfix at home, but I'm glad I made the move. Not only do I have less twiddling to do on my own time, but when my stuff was all torn down and shipped here from Phoenix, I didn't lost anything. Good luck! - Bob
From: "Bob George" <mailings02@ttlexceeded.com>
"Richard Goodman" <r.goodman@11harvard.com> wrote:
My home DSL has been down for over 48 hours and it appears that it will be down for several more days. (I spare you the gory details). I am concerned about losing mail coming into various domains at that site.
Hey, I wouldn't mind a few of the sort-of-gory details. I have recently installed Speakeasy DSL and am thinking of running my own DNS and mail. I am willing to have a bit of trouble in order to learn how it works, but is it common for DSL to "go down"? Why is that? It goes through the phone line, and in nearly half a century on this planet I don't remember the phone not working since a tornado went by a block from our house when I was about ten years old. Of course, except for dial-up internet access, I can easily go for months without making a phone call. If it did go down I might not notice. Was it a problem on their end, or did you drop a brick on your modem?
I was in a similar situation not too long ago when prolonged outages, address changes and ISP port blocking became too much to deal with. Losing my mail was fine, but the family went nuts when I was on the road and we had an outage.
What's up with that? DSL or TV cable? Why outage?
I wound up moving my SMTP to a $2/mo. hosting plan that will queue mail for me until I poll using POP via fetchmail. I've permanently pointed my MX record there, as their reliability is far better than mine (residential cable).
Wow! Where did you find mail hosting for $2/mo? -- Keith
On Thursday 02 October 2003 00:02, Keith Wright wrote:
Hey, I wouldn't mind a few of the sort-of-gory details. I have recently installed Speakeasy DSL and am thinking of running my own DNS and mail. I am willing to have a bit of trouble in order to learn how it works, but is it common for DSL to "go down"? Why is that?
usually it depends on the provider ... i had shitty service with verizon and ive heard similar horror stories with earthlink ... speakeasy is the only one ive never heard anything bad about ... but thats probably why they cost a little bit more than everyone else :) -mike
Mike Frysinger <vapier@gentoo.org> writes:
usually it depends on the provider ... i had shitty service with verizon and ive heard similar horror stories with earthlink ... speakeasy is the only one ive never heard anything bad about ... but thats probably why they cost a little bit more than everyone else :)
Speakeasy is excellent. I've had them at my current apartment for 1 1/2 years, and there were only two (very short) outages. They don't mind if you share your connection via wireless, you get static ip addresses, they don't block ports, let you run servers, etc. Very good ISP. -- Josh Huber
"Keith Wright" <kwright@free-comp-shop.com> wrote:
[...] Hey, I wouldn't mind a few of the sort-of-gory details.
I'll respond to my bits!
I have recently installed Speakeasy DSL and am thinking of running my own DNS and mail.
I have no first-hand experience with DSL, but I would expect it's shares a lot of characteristics (in terms of terms-of-services, reliability) with other residential services. There's usually no guarantee of stellar reliability, and many don't support (or allow) "servers". That's an important consideration if your mail is important to you.
I am willing to have a bit of trouble in order to learn how it works, but is it common for DSL to "go down"? Why is that?
Again, I don't use DSL, but I have had plenty of outages both here and back in Phoenix that have nothing to do with the unerlying technology. Your ISP can have routing problems to the rest of the world, their equipment or DNS servers can be misconfigured, and of course denial-of-service attacks can take their toll. Here, I've become much more familiar with power-related problems. Not "common" but certainly not business-class either.
[...] Losing my mail was fine, but the family went nuts when I
was on the road and we had an outage.
What's up with that? DSL or TV cable? Why outage?
I wound up moving my SMTP to a $2/mo. hosting plan that will queue mail for me until I poll using POP via fetchmail. I've
In the example I cited, it was HFC (cable). The reasons for the outage varied from power outages (my end and theirs), changes in service (it's residential -- they change it when they want i.e. sudden port blocking) and those mysterious 3 hour drops that I never could pin down exactly. permanently
pointed my MX record there, as their reliability is far better than mine (residential cable).
Wow! Where did you find mail hosting for $2/mo?
There are no doubt others, but I went with bmhost.com. It's nothing stellar if you're after hosting for a big website, but you get 10M storage and connectivity seems good. I poll it every 5 minutes via fetchmail and sort things out on my server. They're not particulary linux-aware though, but the servers are RedHat. You can get (chroot) ssh access. The important thing is I can survive an outage on my end without losing mail, and they offer webmail for those situations. Tell 'em I sent ya (ttlexceeded.com is the domain I host there) and we can probably pool to save a few bucks. I paid dyndns.org the $30 one-time fee for "custom" DNS, so I've got my DNS set up there, with dynamically updated entries for my home system (updated via the dhclient scripts), but nice, stable entries pointing at bmhost for MX and web addresses. The only thing I can't do is set up anti-UCE on my SMTP server though bmhost offers some, and they do provide anti-virus scanning standard, so I don't get the glamor of fighting those on my own... which I've found gives me a lot more free time. Of course, none of this applies if you're running a business on your home system. - Bob
"Keith Wright" <kwright@free-comp-shop.com> wrote:
[...] Wow! Where did you find mail hosting for $2/mo?
I replied earlier that I'm using bmhost.com, which provides SMTP and web hosting. I noticed today that dyndns.org has made some changes (good and bad): Custom DNS WAS $30 one-time, but is now an annual charge. They now can do domain registration for $15/yr. for .com (I may transfer mine). More relevant to this thread, they now offer two mail-related services that might be of interest to those on this list running email servers, but who DON'T have a handy second server available for backup, or whose ISP blocks inbound TCP port 25. Quoting from their web page: --- cut here --- cut here --- MailHopSM Relay allows you to relay incoming mail to a mail server on a non-standard port. This is quite useful for those who want to run a mail server but are stuck with an ISP that blocks port 25, the standard mail port. [...] ($24.95/domain) MailHopSM Backup MX provides a secondary mail server to hold your e-mail for you should your main e-mail server go offline for any reason. Once you go back online, your mail will be delivered to you. Our MailHopSM server will attempt delivery more frequently than most mail servers, and will also hold onto your e-mail for twice as long. ($39.95/domain) --- cut here --- cut here --- Both are more expensive than my cheapie hosting outfit, but do allow you to run a "real" SMTP server if so desired. I'm not affiliated with either bmhost.com or dyndns.org, but am a satisfied customer of both. I just wanted to pass this info on to others experiencing the same challenges running home servers that I have. - Bob
participants (7)
-
Bob George -
Josh Huber -
Karl Hiramoto -
Keith Wright -
Mike Clark -
Mike Frysinger -
Richard Goodman