Issue System cannot fork?

[tkalfaoglu]

Server operating system version

AlmaLinux

Plesk version and microupdate number

Obsidian

I am seeing this error in the logs and outgoing mail is not accepted:

Jun 07 14:52:49 server.xxx.com postfix/smtpd[16847]: warning: milter inet:127.0.0.1:12768: can't read SMFIC_OPTNEG reply packet header: Connection timed out
Jun 07 14:52:49 server.xxx.com postfix/smtpd[16847]: warning: milter inet:127.0.0.1:12768: read error in initial handshake
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user.control/run-user-10296.mount.wants': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.wants': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user.control/run-user-10296.mount.requires': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.requires': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user.control/run-user-10296.mount.d': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.d': Permission denied
Jun 07 14:52:50 server.xxx.com dk_check[18959]: 2565841415D6: DKIM Feed: No signature
Jun 07 14:52:50 server.xxx.com postfix/master[19000]: fatal: open lock file /var/lib/postfix/master.lock: unable to set exclusive lock: Resource temporarily unavailable
Jun 07 14:52:51 server.xxx.com postfix/master[13133]: warning: master_spawn: fork: Resource temporarily unavailable -- throttling
Jun 07 14:52:51 server.xxx.com postfix/smtpd[19024]: warning: hostname bl2.41.215.215.184.dynamic.dsl.cvmultimedia.cv does not resolve to address 41.215.215.184: Name or service not known
Jun 07 14:52:51 server.xxx.com postfix/master[18999]: fatal: daemon initialization failure
Jun 07 14:52:52 server.xxx.com postfix/smtpd[14078]: warning: milter inet:127.0.0.1:12768: can't read SMFIC_OPTNEG reply packet header: Connection timed out
Jun 07 14:52:52 server.xxx.com postfix/smtpd[14078]: warning: milter inet:127.0.0.1:12768: read error in initial handshake
Jun 07 14:52:52 server.xxx.com magicspam-plesk[14860]: lm_file.c:267 ERROR lm_read(): Connection reset by peer
Jun 07 14:52:52 server.xxx.com magicspam-plesk[14860]: msd-packet.c:742 ERROR msd_response_packet_receive(): Failed to read daemon response packet: Connection reset by peer
Jun 07 14:52:52 server.xxx.com magicspam-plesk[14860]: smtp-handlers.c:712 ERROR __process_smtp_state_to_response(): Failed to receive daemon response packet from the server
Jun 07 14:52:52 server.xxx.com postfix/postfix-script[19035]: fatal: mail system startup failed
Jun 07 14:52:52 server.xxx.com dovecot[61911]: master: Error: service(imap): fork() failed: Resource temporarily unavailable (ulimit -u 255893 reached?)
Jun 07 14:52:52 server.xxx.com dovecot[61911]: master: Error: service(imap): command startup failed, throttling for 2.000 secs
Jun 07 14:52:53 server.xxx.com postfix/master[13133]: warning: master_spawn: fork: Resource temporarily unavailable -- throttling

The customer sees this error:
<[email protected]>
Command died with status 1:
"/usr/lib64/plesk-9.0/postfix-local". Command output: fork_execv_pipes: fork() failed System error 11: Resource temporarily unavailable

I thought I had increase all ulimit's etc enough.. So why does this happen I wonder?
Thanks! -t

 

[IgorG]

A huge amount of mail to deliver. The operating system imposes a limit on the concurrent threads/processes.

Try to implement the solution from there:
Postfix: limit outgoing mail throttling - Deimos.fr / Bloc Notes Informatique

If the issue persists, consider adjusting the systemd limits as described here:
spawn: fork: Resource temporarily unavailable -- throttling

 

[tkalfaoglu]

Thank you very much -- I made those changes to the system.

I know this is unrelated, but any ideas what to do about these? They repeat at every 10 seconds in the logs..

Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.wants': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user.control/run-user-10296.mount.requires': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.requires': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user.control/run-user-10296.mount.d': Permission denied
Jun 07 14:52:50 server.xxx.com systemd[10362]: Failed to canonicalize path '/usr/local/psa/admin/.config/systemd/user/run-user-10296.mount.d': Permission denied

 

[tkalfaoglu]

Btw, those paths do not exist.. there is almost nothing under the .config directory, except the configstore directory

 

[ciB]

A huge amount of mail to deliver. The operating system imposes a limit on the concurrent threads/processes.

Try to implement the solution from there:
Postfix: limit outgoing mail throttling - Deimos.fr / Bloc Notes Informatique

If the issue persists, consider adjusting the systemd limits as described here:
spawn: fork: Resource temporarily unavailable -- throttling

Sorry coming back to a relatively old topic, but I am having the same issue when sending out a big amount of mails at the same time. Where exactly to I implement those settings in Plesk?

 

[Peter Debik]

@ciB The settings described in the external documentation need to be applied to the Postfix configuration file(s) on the operating system level. They cannot be applied through the Plesk GUI.

 

[tkalfaoglu]

Interestingly I have the same issues again.. Different OS (Rocky linux) even..

Error: osdetect failed: runtime: failed to create new OS thread (have 2 already; errno=11) runtime: may need to increase max user processes (ulimit -u) fatal error: newosproc runtime stack: runtime.throw({0x4edde2?, 0x7fffe86b1b70?}) /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/panic.go:1047 +0x5d fp=0x7fffe86b1b48 sp=0x7fffe86b1b18 pc=0x432a5d runtime.newosproc(0xc00004a000) /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/os_linux.go:195 +0x176 fp=0x7fffe86b1bb0 sp=0x7fffe86b1b48 pc=0x42f4f6 runtime.newm1(0xc00004a000) /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/proc.go:2254 +0xcf fp=0x7fffe86b1bf0 sp=0x7fffe86b1bb0 pc=0x4393cf runtime.newm(0x43dcd1?, 0x0, 0x5b5600?) /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/proc.go:2229 +0x12c fp=0x7fffe86b1c20 sp=0x7fffe86b1bf0 pc=0x43928c runtime.main.func1() /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/proc.go:171 +0x29 fp=0x7fffe86b1c48 sp=0x7fffe86b1c20 pc=0x45afc9 runtime.systemstack() /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/asm_amd64.s:496 +0x49 fp=0x7fffe86b1c50 sp=0x7fffe86b1c48 pc=0x45eaa9 goroutine 1 [running]: runtime.systemstack_switch() /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/asm_amd64.s:463 fp=0xc000046780 sp=0xc000046778 pc=0x45ea40 runtime.main() /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/proc.go:170 +0x6d fp=0xc0000467e0 sp=0xc000046780 pc=0x4351ed runtime.goexit() /var/jenkins/workspace/PLESK/plesk-aws-bootstrap/buck-out/gen/golang/src/runtime/asm_amd64.s:1598 +0x1 fp=0xc0000467e8 sp=0xc0000467e0 pc=0x460b21

 

[Peter Debik]

@tkalfaoglu Looks like a different issue: Not enough concurrent threads possible. "may need to increase max user processes (ulimit -u)". A limit of the operating system settings.

 

[ciB]

Unfortunately those changes do not seem to add a delay to messages - checking the logs, when a client is sending a message to ~100 addresses, they are all sent out at once (logs only show ms in between) - which causes the fork error. Any other ideas?