Re: 2.3.17 broken on CentOS8 / bug
> "Aki" == Aki Tuomi writes: Aki> You are correct that the problem is not fully fixed yet. It, Aki> however, only affects practically cases where you do doveadm -c Aki> /path Thanks for the update. Aki> We will fix it properly in a future release, now it has been Aki> fixed to work as it used to before, so no new regression is Aki> introduced. As long as no one trips over this issue with too long certs some other way. >> On 03/11/2021 14:54 John Stoffel wrote: >> >> >> > "Aki" == Aki Tuomi writes: >> Aki> This issue is now fixed for Dovecot on master with Aki> https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch >> >> Looking at the patch, I've got a couple of comments. >> >> 1. Even your added comment says this issue could still happen is >> doveadm reads the config setting through doveconf, instead of the >> config socket. To me that smells like the problem isn't really where >> you patched it, but more in the parsing of options in doveadm. >> >> 2. This is much more bike-shedding, but you have the following: >> >> -if (input->module != NULL || input->extra_modules != NULL) { >> +if ((service->flags & MASTER_SERVICE_FLAG_DISABLE_SSL_SET) == >> 0 && >> + (input->module != NULL || input->extra_modules != NULL)) { >> >> And I would think that the last line would be more readable with: >> >> (input->module || input->extra_modules)) { >> >> The != NULL test just seems really redundant. I haven't looked at the >> rest of the main.c to see if this pattern is repeated all over the >> place or not. >> >> John >> >> Aki> and for pigeonhole master with >> Aki> https://github.com/dovecot/pigeonhole/commit/29750ba54c20eea0afd4ca436ddc1325723ce93f.patch >> Aki> Regards, Aki> Aki >> >> >> On 01/11/2021 08:38 Aki Tuomi wrote: >> >> >> >> >> >> Hi all! >> >> >> >> We are looking into this issue. >> >> >> >> Aki >> >> >> >> > On 30/10/2021 19:36 TG Servers wrote: >> >> > >> >> > >> >> > Thanks Robert, I read that. I will also wait for a patch and stay >> >> > >> >> > Cheers >> >> > >> >> > >> >> > On 30/10/2021 12:59, Robert Nowotny wrote: >> >> > >> >> > > the reason is : >> >> > > >> >> > > ssl_ca = > >> > > >> >> > > if "ca-bundle.crt"is too big, You will get that error. >> >> > > this should be fixed, but as a workaround You might pull out the >> >> > > certificates You need. >> >> > > I personally wait for the patch and stay at 2.3.16 for the time >> >> > > beeing. >> >> > > >> >> > > yours sincerely >> >> > > Robert >> >> > > >> >> > > >> >> > > >> >> > > Am 30.10.2021 um 10:34 schrieb TG Servers: >> >> > > >> >> > > > Hello, >> >> > > > >> >> > > > tonight my dovecot upgraded to 2.3.17 and completely broke on >> >> > > > recent CentOS 8 installation. >> >> > > > >> >> > > > I found the service in status >> >> > > > >> >> > > > [root@riot ~]# systemctl status dovecot >> >> > > > ● dovecot.service - Dovecot IMAP/POP3 email server >> >> > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; >> >> > > > vendor preset: disabled) >> >> > > > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 >> >> > > > CEST; 58s ago >> >> > > > Docs: man:dovecot(1) >> >> > > > https://doc.dovecot.org/ >> >> > > > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, >> >> > > > status=89) >> >> > > > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript >> >> > > > (code=exited, status=0/SUCCESS) >> >> > > > Main PID: 1515 (code=exited, status=89) >> >> > > > >> >> > > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot >> >> > > > IMAP/POP3 email server... >> >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >> >> > > > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too >> >> > > > long >> >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: >> >> > > > managesieve-login: dump-capability process returned 89 >> >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >> >> > > > execvp(/usr/sbin/dovecot) failed: Argument list too long >> >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: >> >> > > > Main process exited, code=exited, status=89/n/a >> >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: >> >> > > > Failed with result 'exit-code'. >> >> > > > Oct 30 09:59:11 riot..com systemd[1]: Failed to start >> >> > > > Dovecot IMAP/POP3 email server. >> >> > > > >> >> > > > This seems to be like a bug as no configuration was changed by me >> >> > > > in the middle of the night. >> >> > > > I recall there were similar errors/bug reports in the past were it >> >> > > > seemed it was managesieve but wasn't, people had some >> >> > > > misconfigurations in the dovecot.conf. I did not change my >> >> > > > dovecot.conf since April. >> >> > > > But maybe here it is a pigeonhole issue. >> >> > > > >> >> > > > As I did not find any reason for it I changed the repo and
Re: 2.3.17 broken on CentOS8 / bug
> Op 3 nov. 2021 om 19:14 heeft Elise het volgende > geschreven: > > Aki Tuomi: >> This issue is now fixed for Dovecot on master with >> https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch > > Can someone hint us how we should process this fix (sorry, blond)? Patch + compile? > > /elise >
Re: 2.3.17 broken on CentOS8 / bug
Aki Tuomi: This issue is now fixed for Dovecot on master with https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch Can someone hint us how we should process this fix (sorry, blond)? /elise
Re: 2.3.17 broken on CentOS8 / bug
You are correct that the problem is not fully fixed yet. It, however, only affects practically cases where you do doveadm -c /path We will fix it properly in a future release, now it has been fixed to work as it used to before, so no new regression is introduced. Aki > On 03/11/2021 14:54 John Stoffel wrote: > > > > "Aki" == Aki Tuomi writes: > > Aki> This issue is now fixed for Dovecot on master with > Aki> https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch > > Looking at the patch, I've got a couple of comments. > > 1. Even your added comment says this issue could still happen is >doveadm reads the config setting through doveconf, instead of the >config socket. To me that smells like the problem isn't really where >you patched it, but more in the parsing of options in doveadm. > > 2. This is much more bike-shedding, but you have the following: > > - if (input->module != NULL || input->extra_modules != NULL) { > + if ((service->flags & MASTER_SERVICE_FLAG_DISABLE_SSL_SET) == > 0 && > + (input->module != NULL || input->extra_modules != NULL)) { > > And I would think that the last line would be more readable with: > > (input->module || input->extra_modules)) { > > The != NULL test just seems really redundant. I haven't looked at the > rest of the main.c to see if this pattern is repeated all over the > place or not. > > John > > > Aki> and for pigeonhole master with > > Aki> > https://github.com/dovecot/pigeonhole/commit/29750ba54c20eea0afd4ca436ddc1325723ce93f.patch > > Aki> Regards, > Aki> Aki > > >> On 01/11/2021 08:38 Aki Tuomi wrote: > >> > >> > >> Hi all! > >> > >> We are looking into this issue. > >> > >> Aki > >> > >> > On 30/10/2021 19:36 TG Servers wrote: > >> > > >> > > >> > Thanks Robert, I read that. I will also wait for a patch and stay > >> > > >> > Cheers > >> > > >> > > >> > On 30/10/2021 12:59, Robert Nowotny wrote: > >> > > >> > > the reason is : > >> > > > >> > > ssl_ca = >> > > > >> > > if "ca-bundle.crt"is too big, You will get that error. > >> > > this should be fixed, but as a workaround You might pull out the > >> > > certificates You need. > >> > > I personally wait for the patch and stay at 2.3.16 for the time > >> > > beeing. > >> > > > >> > > yours sincerely > >> > > Robert > >> > > > >> > > > >> > > > >> > > Am 30.10.2021 um 10:34 schrieb TG Servers: > >> > > > >> > > > Hello, > >> > > > > >> > > > tonight my dovecot upgraded to 2.3.17 and completely broke on > >> > > > recent CentOS 8 installation. > >> > > > > >> > > > I found the service in status > >> > > > > >> > > > [root@riot ~]# systemctl status dovecot > >> > > > ● dovecot.service - Dovecot IMAP/POP3 email server > >> > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; > >> > > > vendor preset: disabled) > >> > > > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 > >> > > > CEST; 58s ago > >> > > > Docs: man:dovecot(1) > >> > > > https://doc.dovecot.org/ > >> > > > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, > >> > > > status=89) > >> > > > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript > >> > > > (code=exited, status=0/SUCCESS) > >> > > > Main PID: 1515 (code=exited, status=89) > >> > > > > >> > > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot > >> > > > IMAP/POP3 email server... > >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > >> > > > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too > >> > > > long > >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: > >> > > > managesieve-login: dump-capability process returned 89 > >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > >> > > > execvp(/usr/sbin/dovecot) failed: Argument list too long > >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main > >> > > > process exited, code=exited, status=89/n/a > >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: > >> > > > Failed with result 'exit-code'. > >> > > > Oct 30 09:59:11 riot..com systemd[1]: Failed to start > >> > > > Dovecot IMAP/POP3 email server. > >> > > > > >> > > > This seems to be like a bug as no configuration was changed by me > >> > > > in the middle of the night. > >> > > > I recall there were similar errors/bug reports in the past were it > >> > > > seemed it was managesieve but wasn't, people had some > >> > > > misconfigurations in the dovecot.conf. I did not change my > >> > > > dovecot.conf since April. > >> > > > But maybe here it is a pigeonhole issue. > >> > > > > >> > > > As I did not find any reason for it I changed the repo and > >> > > > downgraded to 2.3.16-2 now and it runs without any flaws, like all > >> > > > the time before. I had no time to investigate this any longer thand > >> > > > 2 hours with 2.3.17 installed as this is a production server and I
Re: 2.3.17 broken on CentOS8 / bug
> "Aki" == Aki Tuomi writes: Aki> This issue is now fixed for Dovecot on master with Aki> https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch Looking at the patch, I've got a couple of comments. 1. Even your added comment says this issue could still happen is doveadm reads the config setting through doveconf, instead of the config socket. To me that smells like the problem isn't really where you patched it, but more in the parsing of options in doveadm. 2. This is much more bike-shedding, but you have the following: - if (input->module != NULL || input->extra_modules != NULL) { + if ((service->flags & MASTER_SERVICE_FLAG_DISABLE_SSL_SET) == 0 && + (input->module != NULL || input->extra_modules != NULL)) { And I would think that the last line would be more readable with: (input->module || input->extra_modules)) { The != NULL test just seems really redundant. I haven't looked at the rest of the main.c to see if this pattern is repeated all over the place or not. John Aki> and for pigeonhole master with Aki> https://github.com/dovecot/pigeonhole/commit/29750ba54c20eea0afd4ca436ddc1325723ce93f.patch Aki> Regards, Aki> Aki >> On 01/11/2021 08:38 Aki Tuomi wrote: >> >> >> Hi all! >> >> We are looking into this issue. >> >> Aki >> >> > On 30/10/2021 19:36 TG Servers wrote: >> > >> > >> > Thanks Robert, I read that. I will also wait for a patch and stay >> > >> > Cheers >> > >> > >> > On 30/10/2021 12:59, Robert Nowotny wrote: >> > >> > > the reason is : >> > > >> > > ssl_ca = > > > >> > > if "ca-bundle.crt"is too big, You will get that error. >> > > this should be fixed, but as a workaround You might pull out the >> > > certificates You need. >> > > I personally wait for the patch and stay at 2.3.16 for the time beeing. >> > > >> > > yours sincerely >> > > Robert >> > > >> > > >> > > >> > > Am 30.10.2021 um 10:34 schrieb TG Servers: >> > > >> > > > Hello, >> > > > >> > > > tonight my dovecot upgraded to 2.3.17 and completely broke on recent >> > > > CentOS 8 installation. >> > > > >> > > > I found the service in status >> > > > >> > > > [root@riot ~]# systemctl status dovecot >> > > > ● dovecot.service - Dovecot IMAP/POP3 email server >> > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; >> > > > vendor preset: disabled) >> > > > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 >> > > > CEST; 58s ago >> > > > Docs: man:dovecot(1) >> > > > https://doc.dovecot.org/ >> > > > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) >> > > > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript >> > > > (code=exited, status=0/SUCCESS) >> > > > Main PID: 1515 (code=exited, status=89) >> > > > >> > > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot >> > > > IMAP/POP3 email server... >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >> > > > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: >> > > > managesieve-login: dump-capability process returned 89 >> > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >> > > > execvp(/usr/sbin/dovecot) failed: Argument list too long >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main >> > > > process exited, code=exited, status=89/n/a >> > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed >> > > > with result 'exit-code'. >> > > > Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot >> > > > IMAP/POP3 email server. >> > > > >> > > > This seems to be like a bug as no configuration was changed by me in >> > > > the middle of the night. >> > > > I recall there were similar errors/bug reports in the past were it >> > > > seemed it was managesieve but wasn't, people had some >> > > > misconfigurations in the dovecot.conf. I did not change my >> > > > dovecot.conf since April. >> > > > But maybe here it is a pigeonhole issue. >> > > > >> > > > As I did not find any reason for it I changed the repo and downgraded >> > > > to 2.3.16-2 now and it runs without any flaws, like all the time >> > > > before. I had no time to investigate this any longer thand 2 hours >> > > > with 2.3.17 installed as this is a production server and I need the >> > > > email access. I also did not find anything adressable in the logs. >> > > > >> > > > [root@riot dovecot]# systemctl status dovecot >> > > > ● dovecot.service - Dovecot IMAP/POP3 email server >> > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; >> > > > vendor preset: disabled) >> > > > Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago >> > > > Docs: man:dovecot(1) >> > > > https://doc.dovecot.org/ >> > > > Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript >> > > > (code=exited, status=0/SUCCESS) >> > > >
Re: 2.3.17 broken on CentOS8 / bug
This issue is now fixed for Dovecot on master with https://github.com/dovecot/core/compare/ca2237e%5E..6fff8d5.patch and for pigeonhole master with https://github.com/dovecot/pigeonhole/commit/29750ba54c20eea0afd4ca436ddc1325723ce93f.patch Regards, Aki > On 01/11/2021 08:38 Aki Tuomi wrote: > > > Hi all! > > We are looking into this issue. > > Aki > > > On 30/10/2021 19:36 TG Servers wrote: > > > > > > Thanks Robert, I read that. I will also wait for a patch and stay > > > > Cheers > > > > > > On 30/10/2021 12:59, Robert Nowotny wrote: > > > > > the reason is : > > > > > > ssl_ca = > > > > > if "ca-bundle.crt"is too big, You will get that error. > > > this should be fixed, but as a workaround You might pull out the > > > certificates You need. > > > I personally wait for the patch and stay at 2.3.16 for the time beeing. > > > > > > yours sincerely > > > Robert > > > > > > > > > > > > Am 30.10.2021 um 10:34 schrieb TG Servers: > > > > > > > Hello, > > > > > > > > tonight my dovecot upgraded to 2.3.17 and completely broke on recent > > > > CentOS 8 installation. > > > > > > > > I found the service in status > > > > > > > > [root@riot ~]# systemctl status dovecot > > > > ● dovecot.service - Dovecot IMAP/POP3 email server > > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; > > > > vendor preset: disabled) > > > > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; > > > > 58s ago > > > > Docs: man:dovecot(1) > > > > https://doc.dovecot.org/ > > > > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) > > > > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript > > > > (code=exited, status=0/SUCCESS) > > > > Main PID: 1515 (code=exited, status=89) > > > > > > > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot > > > > IMAP/POP3 email server... > > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > > > > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long > > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: > > > > managesieve-login: dump-capability process returned 89 > > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > > > > execvp(/usr/sbin/dovecot) failed: Argument list too long > > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main > > > > process exited, code=exited, status=89/n/a > > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed > > > > with result 'exit-code'. > > > > Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot > > > > IMAP/POP3 email server. > > > > > > > > This seems to be like a bug as no configuration was changed by me in > > > > the middle of the night. > > > > I recall there were similar errors/bug reports in the past were it > > > > seemed it was managesieve but wasn't, people had some misconfigurations > > > > in the dovecot.conf. I did not change my dovecot.conf since April. > > > > But maybe here it is a pigeonhole issue. > > > > > > > > As I did not find any reason for it I changed the repo and downgraded > > > > to 2.3.16-2 now and it runs without any flaws, like all the time > > > > before. I had no time to investigate this any longer thand 2 hours with > > > > 2.3.17 installed as this is a production server and I need the email > > > > access. I also did not find anything adressable in the logs. > > > > > > > > [root@riot dovecot]# systemctl status dovecot > > > > ● dovecot.service - Dovecot IMAP/POP3 email server > > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; > > > > vendor preset: disabled) > > > > Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago > > > > Docs: man:dovecot(1) > > > > https://doc.dovecot.org/ > > > > Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript > > > > (code=exited, status=0/SUCCESS) > > > > Main PID: 32452 (dovecot) > > > > Status: "v2.3.16 (7e2e900c1a) running" > > > > Tasks: 4 (limit: 99912) > > > > Memory: 4.4M > > > > CGroup: /system.slice/dovecot.service > > > > ├─32452 /usr/sbin/dovecot -F > > > > ├─32507 dovecot/anvil > > > > ├─32508 dovecot/log > > > > └─32513 dovecot/config > > > > > > > > Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot > > > > IMAP/POP3 email server... > > > > Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected > > > > permissions for login directory /var/run/dovecot/token-login > > > > Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: > > > > Corrected permissions for login directory /var/run/dovecot/token-login > > > > Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot > > > > v2.3.16 (7e2e900c1a) starting up for imap, lmtp, sieve > > > > Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot > > > > IMAP/POP3 email server. > > > > > > > > > > > > This is the configuration > > > > # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf > > > >
Re: 2.3.17 broken on CentOS8 / bug
Hi all! We are looking into this issue. Aki > On 30/10/2021 19:36 TG Servers wrote: > > > Thanks Robert, I read that. I will also wait for a patch and stay > > Cheers > > > On 30/10/2021 12:59, Robert Nowotny wrote: > > > the reason is : > > > > ssl_ca = > > > if "ca-bundle.crt"is too big, You will get that error. > > this should be fixed, but as a workaround You might pull out the > > certificates You need. > > I personally wait for the patch and stay at 2.3.16 for the time beeing. > > > > yours sincerely > > Robert > > > > > > > > Am 30.10.2021 um 10:34 schrieb TG Servers: > > > > > Hello, > > > > > > tonight my dovecot upgraded to 2.3.17 and completely broke on recent > > > CentOS 8 installation. > > > > > > I found the service in status > > > > > > [root@riot ~]# systemctl status dovecot > > > ● dovecot.service - Dovecot IMAP/POP3 email server > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor > > > preset: disabled) > > > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; > > > 58s ago > > > Docs: man:dovecot(1) > > > https://doc.dovecot.org/ > > > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) > > > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript > > > (code=exited, status=0/SUCCESS) > > > Main PID: 1515 (code=exited, status=89) > > > > > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 > > > email server... > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > > > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: > > > managesieve-login: dump-capability process returned 89 > > > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > > > execvp(/usr/sbin/dovecot) failed: Argument list too long > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main > > > process exited, code=exited, status=89/n/a > > > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed > > > with result 'exit-code'. > > > Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot > > > IMAP/POP3 email server. > > > > > > This seems to be like a bug as no configuration was changed by me in the > > > middle of the night. > > > I recall there were similar errors/bug reports in the past were it > > > seemed it was managesieve but wasn't, people had some misconfigurations > > > in the dovecot.conf. I did not change my dovecot.conf since April. > > > But maybe here it is a pigeonhole issue. > > > > > > As I did not find any reason for it I changed the repo and downgraded to > > > 2.3.16-2 now and it runs without any flaws, like all the time before. I > > > had no time to investigate this any longer thand 2 hours with 2.3.17 > > > installed as this is a production server and I need the email access. I > > > also did not find anything adressable in the logs. > > > > > > [root@riot dovecot]# systemctl status dovecot > > > ● dovecot.service - Dovecot IMAP/POP3 email server > > > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor > > > preset: disabled) > > > Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago > > > Docs: man:dovecot(1) > > > https://doc.dovecot.org/ > > > Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript > > > (code=exited, status=0/SUCCESS) > > > Main PID: 32452 (dovecot) > > > Status: "v2.3.16 (7e2e900c1a) running" > > > Tasks: 4 (limit: 99912) > > > Memory: 4.4M > > > CGroup: /system.slice/dovecot.service > > > ├─32452 /usr/sbin/dovecot -F > > > ├─32507 dovecot/anvil > > > ├─32508 dovecot/log > > > └─32513 dovecot/config > > > > > > Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 > > > email server... > > > Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected > > > permissions for login directory /var/run/dovecot/token-login > > > Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: > > > Corrected permissions for login directory /var/run/dovecot/token-login > > > Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot > > > v2.3.16 (7e2e900c1a) starting up for imap, lmtp, sieve > > > Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 > > > email server. > > > > > > > > > This is the configuration > > > # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf > > > # Pigeonhole version 0.5.16 (09c29328) > > > # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 > > > (Electric Cheetah) > > > # Hostname: riot..com > > > auth_mechanisms = plain login > > > auth_verbose = yes > > > listen = * > > > mail_gid = vmail > > > mail_home = /var/vmail/mailboxes/%d/%n > > > mail_location = maildir:~/mail:LAYOUT=fs > > > mail_plugins = " quota fts fts_solr" > > > mail_privileged_group = vmail > > > mail_uid = vmail > > > managesieve_notify_capability = mailto > > > manag
Re: 2.3.17 broken on CentOS8 / bug
Thanks Robert, I read that. I will also wait for a patch and stay Cheers On 30/10/2021 12:59, Robert Nowotny wrote: the reason is : ssl_ca = if "ca-bundle.crt" is too big, You will get that error. this should be fixed, but as a workaround You might pull out the certificates You need. I personally wait for the patch and stay at 2.3.16 for the time beeing. yours sincerely Robert Am 30.10.2021 um 10:34 schrieb TG Servers: Hello, tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 installation. I found the service in status [root@riot ~]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 1515 (code=exited, status=89) Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: managesieve-login: dump-capability process returned 89 Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/sbin/dovecot) failed: Argument list too long Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process exited, code=exited, status=89/n/a Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with result 'exit-code'. Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot IMAP/POP3 email server. This seems to be like a bug as no configuration was changed by me in the middle of the night. I recall there were similar errors/bug reports in the past were it seemed it was managesieve but wasn't, people had some misconfigurations in the dovecot.conf. I did not change my dovecot.conf since April. But maybe here it is a pigeonhole issue. As I did not find any reason for it I changed the repo and downgraded to 2.3.16-2 now and it runs without any flaws, like all the time before. I had no time to investigate this any longer thand 2 hours with 2.3.17 installed as this is a production server and I need the email access. I also did not find anything adressable in the logs. [root@riot dovecot]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 32452 (dovecot) Status: "v2.3.16 (7e2e900c1a) running" Tasks: 4 (limit: 99912) Memory: 4.4M CGroup: /system.slice/dovecot.service ├─32452 /usr/sbin/dovecot -F ├─32507 dovecot/anvil ├─32508 dovecot/log └─32513 dovecot/config Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 (7e2e900c1a) starting up for imap, lmtp, sieve Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 email server. This is the configuration # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf
Re: 2.3.17 broken on CentOS8 / bug
On 30/10/2021 13:04, William Edwards wrote: Op 30 okt. 2021 om 12:10 heeft TG Servers het volgende geschreven: Thanks for your reply William. But the only thing I found in the meanwhile about this issue is that when the ca-bundles files is too "big" it does not work anymore. And if this file is shortened to one entry it will work, someone seems to have tested this. This is no fix, it is a bug that has to be fixed by dovecot from my pov. A fix and a bug are not mutually exclusive :) You are surely right on that one :) But in this case it is also no fix for me, I would either have to use a butchered file just for dovecot or change the file for all. It is no big issue though, I think it is obviously this should be fixed in one of the next releases The ca-bundles file is used by countless applications without any issues, it is used by 2.3.16 without any issues. There should be no special treatment for a single application necessary. On 30/10/2021 11:35, William Edwards wrote: Op 30 okt. 2021 om 10:35 heeft TG Servers het volgende geschreven: Hello, tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 installation. I found the service in status [root@riot ~]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 1515 (code=exited, status=89) Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: managesieve-login: dump-capability process returned 89 Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/sbin/dovecot) failed: Argument list too long Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process exited, code=exited, status=89/n/a Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with result 'exit-code'. Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot IMAP/POP3 email server. Please check the archive. If I’m not mistaken, the same issue + possible solution was posted on the mailing list yesterday. This seems to be like a bug as no configuration was changed by me in the middle of the night. I recall there were similar errors/bug reports in the past were it seemed it was managesieve but wasn't, people had some misconfigurations in the dovecot.conf. I did not change my dovecot.conf since April. But maybe here it is a pigeonhole issue. As I did not find any reason for it I changed the repo and downgraded to 2.3.16-2 now and it runs without any flaws, like all the time before. I had no time to investigate this any longer thand 2 hours with 2.3.17 installed as this is a production server and I need the email access. I also
Re: 2.3.17 broken on CentOS8 / bug
> Op 30 okt. 2021 om 12:10 heeft TG Servers het volgende > geschreven: > > Thanks for your reply William. > > But the only thing I found in the meanwhile about this issue is that when the > ca-bundles files is too "big" it does not work anymore. And if this file is > shortened to one entry it will work, someone seems to have tested this. > This is no fix, it is a bug that has to be fixed by dovecot from my pov. A fix and a bug are not mutually exclusive :) > The ca-bundles file is used by countless applications without any issues, it > is used by 2.3.16 without any issues. There should be no special treatment > for a single application necessary. > > > > On 30/10/2021 11:35, William Edwards wrote: >> Op 30 okt. 2021 om 10:35 heeft TG Servers het volgende geschreven: >>> Hello, >>> >>> tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS >>> 8 installation. >>> >>> I found the service in status >>> >>> [root@riot ~]# systemctl status dovecot >>> ● dovecot.service - Dovecot IMAP/POP3 email server >>>Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor >>> preset: disabled) >>>Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; >>> 58s ago >>> Docs: man:dovecot(1) >>>https://doc.dovecot.org/ >>> Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) >>> Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript >>> (code=exited, status=0/SUCCESS) >>> Main PID: 1515 (code=exited, status=89) >>> >>> Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 >>> email server... >>> Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >>> execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long >>> Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: >>> managesieve-login: dump-capability process returned 89 >>> Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >>> execvp(/usr/sbin/dovecot) failed: Argument list too long >>> Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process >>> exited, code=exited, status=89/n/a >>> Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with >>> result 'exit-code'. >>> Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot >>> IMAP/POP3 email server. >> >> Please check the archive. If I’m not mistaken, the same issue + possible >> solution was posted on the mailing list yesterday. >> >>> >>> This seems to be like a bug as no configuration was changed by me in the >>> middle of the night. >>> I recall there were similar errors/bug reports in the past were it seemed >>> it was managesieve but wasn't, people had some misconfigurations in the >>> dovecot.conf. I did not change my dovecot.conf since April. >>> But maybe here it is a pigeonhole issue. >>> >>> As I did not find any reason for it I changed the repo and downgraded to >>> 2.3.16-2 now and it runs without any flaws, like all the time before. I had >>> no time to investigate this any longer thand 2 hours with 2.3.17 installed >>> as this is a production server and I need the email access. I also did not >>> find anything adressable in the logs. >>> >>> [root@riot dovecot]# systemctl status dovecot >>> ● dovecot.service - Dovecot IMAP/POP3 email server >>>Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor >>> preset: disabled) >>>Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago >>> Docs: man:dovecot(1) >>>https://doc.dovecot.org/ >>> Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript >>> (code=exited, status=0/SUCCESS) >>> Main PID: 32452 (dovecot) >>>Status: "v2.3.16 (7e2e900c1a) running" >>> Tasks: 4 (limit: 99912) >>>Memory: 4.4M >>>CGroup: /system.slice/dovecot.service >>>├─32452 /usr/sbin/dovecot -F >>>├─32507 dovecot/anvil >>>├─32508 dovecot/log >>>└─32513 dovecot/config >>> >>> Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 >>> email server... >>> Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected >>> permissions for login directory /var/run/dovecot/token-login >>> Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: >>> Corrected permissions for login directory /var/run/dovecot/token-login >>> Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 >>> (7e2e900c1a) starting up for imap, lmtp, sieve >>> Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 >>> email server. >>> >>> >>> This is the configuration >>> # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf >>> # Pigeonhole version 0.5.16 (09c29328) >>> # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 >>> (Electric Cheetah) >>> # Hostname: riot..com >>> auth_mechanisms = plain login >>> auth_verbose = yes >>> listen = * >>> mail_gid = vmail >>> mail_home = /var/vmail/mailboxes/%d/%n >>> mail_
Re: 2.3.17 broken on CentOS8 / bug
the reason is : ssl_ca = this should be fixed, but as a workaround You might pull out the certificates You need. I personally wait for the patch and stay at 2.3.16 for the time beeing. yours sincerely Robert Am 30.10.2021 um 10:34 schrieb TG Servers: Hello, tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 installation. I found the service in status [root@riot ~]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 1515 (code=exited, status=89) Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: managesieve-login: dump-capability process returned 89 Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/sbin/dovecot) failed: Argument list too long Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process exited, code=exited, status=89/n/a Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with result 'exit-code'. Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot IMAP/POP3 email server. This seems to be like a bug as no configuration was changed by me in the middle of the night. I recall there were similar errors/bug reports in the past were it seemed it was managesieve but wasn't, people had some misconfigurations in the dovecot.conf. I did not change my dovecot.conf since April. But maybe here it is a pigeonhole issue. As I did not find any reason for it I changed the repo and downgraded to 2.3.16-2 now and it runs without any flaws, like all the time before. I had no time to investigate this any longer thand 2 hours with 2.3.17 installed as this is a production server and I need the email access. I also did not find anything adressable in the logs. [root@riot dovecot]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 32452 (dovecot) Status: "v2.3.16 (7e2e900c1a) running" Tasks: 4 (limit: 99912) Memory: 4.4M CGroup: /system.slice/dovecot.service ├─32452 /usr/sbin/dovecot -F ├─32507 dovecot/anvil ├─32508 dovecot/log └─32513 dovecot/config Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 (7e2e900c1a) starting up for imap, lmtp, sieve Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 email server. This is the configuration # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf # Pigeonhole version 0.5.16 (09c29328) # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 (Electric Cheetah) # Hostname: riot..com auth_mechanisms = plain login auth_verbose = yes listen = * mail_gid = vmail mail_home = /var/vmail/mailboxes/%d/%n mail_location = maildir:~/mail:LAYOUT=fs mail_plugins = " quota fts fts_solr" mail_privileged_group = vmail mail_uid = vmail managesieve_notify_capability = mailto managesieve_sieve_capability = fileinto reject envelope encoded-character vacation subaddress comparator-i;ascii-numeric relational regex imap4flags copy include variables body enotify environment mailbox date index ihave duplicate mime foreverypart extracttext imapsieve vnd.dovecot.imapsieve namespace inbox { inbox = yes location = mailbox Drafts { auto = subscribe special_use = \Drafts } mailbox Sent { auto = subscribe special_use = \Sent } mailbox Spam { auto = subscribe special_use = \Junk } mailbox Trash { auto = subscribe special_use = \Trash } prefix = separator = . type = private } passdb { args = /etc/dovecot/dovecot-sql.conf driver = sql } plugin { fts = solr fts_autoindex = yes fts_solr = url=http://localhost:/solr/dovecot/ imapsieve_mailbox1_before = file:/var/vmail/sieve/global/learn-spam.sieve imapsieve_mai
Re: 2.3.17 broken on CentOS8 / bug
On October 30, 2021 12:00:40 PM GMT+02:00, TG Servers wrote: >Thanks for your reply William. > > But the only thing I found in the meanwhile about this issue is that when >the ca-bundles files is too "big" it does not work anymore. And if this >file is shortened to one entry it will work, someone seems to have tested >this. > This is no fix, it is a bug that has to be fixed by dovecot from my pov. > The ca-bundles file is used by countless applications without any issues, >it is used by 2.3.16 without any issues. There should be no special >treatment for a single application necessary. Do you use client certs? If not, there is no need to even have ssl_ca set, see https://doc.dovecot.org/configuration_manual/dovecot_ssl_configuration/#id10 > > >On 30/10/2021 11:35, William Edwards wrote: > > >Op 30 okt. 2021 om 10:35 heeft TG Servers het >volgende geschreven: > > >Hello, > > tonight my dovecot upgraded to 2.3.17 and completely broke on recent >CentOS 8 installation. > > I found the service in status > > [root@riot ~]# systemctl status dovecot > ● dovecot.service - Dovecot IMAP/POP3 email server > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor >preset: disabled) > Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; >58s ago > Docs: man:dovecot(1) >https://doc.dovecot.org/ > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript >(code=exited, status=0/SUCCESS) > Main PID: 1515 (code=exited, status=89) > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 >email server... > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: >managesieve-login: dump-capability process returned 89 > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: >execvp(/usr/sbin/dovecot) failed: Argument list too long > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main >process exited, code=exited, status=89/n/a > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed >with result 'exit-code'. > Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot >IMAP/POP3 email server. > > > Please check the archive. If I’m not mistaken, the same issue + possible >solution was posted on the mailing list yesterday. > > > This seems to be like a bug as no configuration was changed by me in the >middle of the night. > I recall there were similar errors/bug reports in the past were it seemed >it was managesieve but wasn't, people had some misconfigurations in the >dovecot.conf. I did not change my dovecot.conf since April. > But maybe here it is a pigeonhole issue. > > As I did not find any reason for it I changed the repo and downgraded to >2.3.16-2 now and it runs without any flaws, like all the time before. I had >no time to investigate this any longer thand 2 hours with 2.3.17 installed >as this is a production server and I need the email access. I also did not >find anything adressable in the logs. > > [root@riot dovecot]# systemctl status dovecot > ● dovecot.service - Dovecot IMAP/POP3 email server > Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor >preset: disabled) > Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago > Docs: man:dovecot(1) >https://doc.dovecot.org/ > Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript >(code=exited, status=0/SUCCESS) > Main PID: 32452 (dovecot) > Status: "v2.3.16 (7e2e900c1a) running" > Tasks: 4 (limit: 99912) > Memory: 4.4M > CGroup: /system.slice/dovecot.service > ├─32452 /usr/sbin/dovecot -F > ├─32507 dovecot/anvil > ├─32508 dovecot/log > └─32513 dovecot/config > > Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 >email server... > Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected >permissions for login directory /var/run/dovecot/token-login > Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: >Corrected permissions for login directory /var/run/dovecot/token-login > Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 >(7e2e900c1a) starting up for imap, lmtp, sieve > Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 >email server. > > > This is the configuration > # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf > # Pigeonhole version 0.5.16 (09c29328) > # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 >(Electric Cheetah) > # Hostname: riot..com > auth_mechanisms = plain login > auth_verbose = yes > listen = * > mail_gid = vmail > mail_home = /var/vmail/mailboxes/%d/%n > mail_location = maildir:~/mail:LAYOUT=fs > mail_plugins = " quota fts fts_solr" > mail_privileged_group = vmail > mail_uid = vmail > managesieve_notify_capability = mailto > managesieve_sieve_capability = fileinto reject envelope encoded-character >vacation suba
Re: 2.3.17 broken on CentOS8 / bug
Thanks for your reply William. But the only thing I found in the meanwhile about this issue is that when the ca-bundles files is too "big" it does not work anymore. And if this file is shortened to one entry it will work, someone seems to have tested this. This is no fix, it is a bug that has to be fixed by dovecot from my pov. The ca-bundles file is used by countless applications without any issues, it is used by 2.3.16 without any issues. There should be no special treatment for a single application necessary. On 30/10/2021 11:35, William Edwards wrote: Op 30 okt. 2021 om 10:35 heeft TG Servers het volgende geschreven: Hello, tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 installation. I found the service in status [root@riot ~]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 1515 (code=exited, status=89) Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: managesieve-login: dump-capability process returned 89 Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/sbin/dovecot) failed: Argument list too long Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process exited, code=exited, status=89/n/a Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with result 'exit-code'. Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot IMAP/POP3 email server. Please check the archive. If I’m not mistaken, the same issue + possible solution was posted on the mailing list yesterday. This seems to be like a bug as no configuration was changed by me in the middle of the night. I recall there were similar errors/bug reports in the past were it seemed it was managesieve but wasn't, people had some misconfigurations in the dovecot.conf. I did not change my dovecot.conf since April. But maybe here it is a pigeonhole issue. As I did not find any reason for it I changed the repo and downgraded to 2.3.16-2 now and it runs without any flaws, like all the time before. I had no time to investigate this any longer thand 2 hours with 2.3.17 installed as this is a production server and I need the email access. I also did not find anything adressable in the logs. [root@riot dovecot]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 32452 (dovecot) Status: "v2.3.16 (7e2e900c1a) running" Tasks: 4 (limit: 99912) Memory: 4.4M CGroup: /system.slice/dovecot.service ├─32452 /usr/sbin/dovecot -F ├─32507 dovecot/anvil ├─32508 dovecot/log └─32513 dovecot/config Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Cor
Re: 2.3.17 broken on CentOS8 / bug
> Op 30 okt. 2021 om 10:35 heeft TG Servers het volgende > geschreven: > > Hello, > > tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 > installation. > > I found the service in status > > [root@riot ~]# systemctl status dovecot > ● dovecot.service - Dovecot IMAP/POP3 email server >Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor > preset: disabled) >Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s > ago > Docs: man:dovecot(1) >https://doc.dovecot.org/ > Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) > Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript > (code=exited, status=0/SUCCESS) > Main PID: 1515 (code=exited, status=89) > > Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 > email server... > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: > managesieve-login: dump-capability process returned 89 > Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: > execvp(/usr/sbin/dovecot) failed: Argument list too long > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process > exited, code=exited, status=89/n/a > Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with > result 'exit-code'. > Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot > IMAP/POP3 email server. Please check the archive. If I’m not mistaken, the same issue + possible solution was posted on the mailing list yesterday. > > This seems to be like a bug as no configuration was changed by me in the > middle of the night. > I recall there were similar errors/bug reports in the past were it seemed it > was managesieve but wasn't, people had some misconfigurations in the > dovecot.conf. I did not change my dovecot.conf since April. > But maybe here it is a pigeonhole issue. > > As I did not find any reason for it I changed the repo and downgraded to > 2.3.16-2 now and it runs without any flaws, like all the time before. I had > no time to investigate this any longer thand 2 hours with 2.3.17 installed as > this is a production server and I need the email access. I also did not find > anything adressable in the logs. > > [root@riot dovecot]# systemctl status dovecot > ● dovecot.service - Dovecot IMAP/POP3 email server >Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor > preset: disabled) >Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago > Docs: man:dovecot(1) >https://doc.dovecot.org/ > Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript > (code=exited, status=0/SUCCESS) > Main PID: 32452 (dovecot) >Status: "v2.3.16 (7e2e900c1a) running" > Tasks: 4 (limit: 99912) >Memory: 4.4M >CGroup: /system.slice/dovecot.service >├─32452 /usr/sbin/dovecot -F >├─32507 dovecot/anvil >├─32508 dovecot/log >└─32513 dovecot/config > > Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 > email server... > Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected > permissions for login directory /var/run/dovecot/token-login > Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: Corrected > permissions for login directory /var/run/dovecot/token-login > Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 > (7e2e900c1a) starting up for imap, lmtp, sieve > Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 email > server. > > > This is the configuration > # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf > # Pigeonhole version 0.5.16 (09c29328) > # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 > (Electric Cheetah) > # Hostname: riot..com > auth_mechanisms = plain login > auth_verbose = yes > listen = * > mail_gid = vmail > mail_home = /var/vmail/mailboxes/%d/%n > mail_location = maildir:~/mail:LAYOUT=fs > mail_plugins = " quota fts fts_solr" > mail_privileged_group = vmail > mail_uid = vmail > managesieve_notify_capability = mailto > managesieve_sieve_capability = fileinto reject envelope encoded-character > vacation subaddress comparator-i;ascii-numeric relational regex imap4flags > copy include variables body enotify environment mailbox date index ihave > duplicate mime foreverypart extracttext imapsieve vnd.dovecot.imapsieve > namespace inbox { > inbox = yes > location = > mailbox Drafts { > auto = subscribe > special_use = \Drafts > } > mailbox Sent { > auto = subscribe > special_use = \Sent > } > mailbox Spam { > auto = subscribe > special_use = \Junk > } > mailbox Trash { > auto = subscribe > special_use = \Trash > } > prefix = > separator = . > type = private > } > passdb { > args = /etc/dovecot/dovecot-sql.con
2.3.17 broken on CentOS8 / bug
Hello, tonight my dovecot upgraded to 2.3.17 and completely broke on recent CentOS 8 installation. I found the service in status [root@riot ~]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sat 2021-10-30 09:59:11 CEST; 58s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 1515 ExecStart=/usr/sbin/dovecot -F (code=exited, status=89) Process: 1429 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 1515 (code=exited, status=89) Oct 30 09:59:10 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/libexec/dovecot/managesieve) failed: Argument list too long Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Error: managesieve-login: dump-capability process returned 89 Oct 30 09:59:11 riot..com dovecot[1515]: doveconf: Fatal: execvp(/usr/sbin/dovecot) failed: Argument list too long Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Main process exited, code=exited, status=89/n/a Oct 30 09:59:11 riot..com systemd[1]: dovecot.service: Failed with result 'exit-code'. Oct 30 09:59:11 riot..com systemd[1]: Failed to start Dovecot IMAP/POP3 email server. This seems to be like a bug as no configuration was changed by me in the middle of the night. I recall there were similar errors/bug reports in the past were it seemed it was managesieve but wasn't, people had some misconfigurations in the dovecot.conf. I did not change my dovecot.conf since April. But maybe here it is a pigeonhole issue. As I did not find any reason for it I changed the repo and downgraded to 2.3.16-2 now and it runs without any flaws, like all the time before. I had no time to investigate this any longer thand 2 hours with 2.3.17 installed as this is a production server and I need the email access. I also did not find anything adressable in the logs. [root@riot dovecot]# systemctl status dovecot ● dovecot.service - Dovecot IMAP/POP3 email server Loaded: loaded (/usr/lib/systemd/system/dovecot.service; enabled; vendor preset: disabled) Active: active (running) since Sat 2021-10-30 10:18:11 CEST; 2s ago Docs: man:dovecot(1) https://doc.dovecot.org/ Process: 32398 ExecStartPre=/usr/libexec/dovecot/prestartscript (code=exited, status=0/SUCCESS) Main PID: 32452 (dovecot) Status: "v2.3.16 (7e2e900c1a) running" Tasks: 4 (limit: 99912) Memory: 4.4M CGroup: /system.slice/dovecot.service ├─32452 /usr/sbin/dovecot -F ├─32507 dovecot/anvil ├─32508 dovecot/log └─32513 dovecot/config Oct 30 10:18:11 riot..com systemd[1]: Starting Dovecot IMAP/POP3 email server... Oct 30 10:18:11 riot..com dovecot[32452]: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Warning: Corrected permissions for login directory /var/run/dovecot/token-login Oct 30 10:18:11 riot..com dovecot[32452]: master: Dovecot v2.3.16 (7e2e900c1a) starting up for imap, lmtp, sieve Oct 30 10:18:11 riot..com systemd[1]: Started Dovecot IMAP/POP3 email server. This is the configuration # 2.3.16 (7e2e900c1a): /etc/dovecot/dovecot.conf # Pigeonhole version 0.5.16 (09c29328) # OS: Linux 4.18.0-305.19.1.el8_4.x86_64 x86_64 AlmaLinux release 8.4 (Electric Cheetah) # Hostname: riot..com auth_mechanisms = plain login auth_verbose = yes listen = * mail_gid = vmail mail_home = /var/vmail/mailboxes/%d/%n mail_location = maildir:~/mail:LAYOUT=fs mail_plugins = " quota fts fts_solr" mail_privileged_group = vmail mail_uid = vmail managesieve_notify_capability = mailto managesieve_sieve_capability = fileinto reject envelope encoded-character vacation subaddress comparator-i;ascii-numeric relational regex imap4flags copy include variables body enotify environment mailbox date index ihave duplicate mime foreverypart extracttext imapsieve vnd.dovecot.imapsieve namespace inbox { inbox = yes location = mailbox Drafts { auto = subscribe special_use = \Drafts } mailbox Sent { auto = subscribe special_use = \Sent