Hi, >> For reference, there was nothing further in the system or postfix >> logs, dmesg, or even by increasing the postfix logging level to the >> best of my ability. > > When the kernel reports "Resource temporarily unavailable", that's > all that Postfix can report without second-guessing the kernel.
Yes, understood. I was more referring to having postfix better indicate what it was trying to do that caused the error, but I guess that's clear to me now that it was trying to change user IDs. >> # ps ax | grep master >> 14352 ? Ss 0:08 /usr/libexec/postfix/master >> 14368 ? Ss 0:00 /usr/libexec/postfix/master > > The presence of two Postfix master processes implies that you are > running two Postfix instances, each with its own queue directory > etc. Yes, I am running two instances of postfix as you've described. >> # lsof -p 14352 -p 14368 | wc -l >> 188 > > Keep in mind that this line count includes the one-line header > (COMMAND PID USER...), the master executable itself, a bunch of > shared libraries (libc, libssl, ...), as well as the per-process > current directory and root directory. Yes, understood. I included that output more for completeness than anything else -- I just increased the maximum processes per user to 260 from 160. Thanks again, Alex
