Slow boot after testing external monitor

[Doug Hutcheson]

I tried plugging an IBM C51 CRT into the external video port on my Thinkpad T61 (running Fedora 15) and rebooted. The display worked OK, but the boot stopped - or appeared to - when the boot progress bar across the bottom of the screen had all turned white. I left it for a minute, but nothing more happened, so I unplugged the monitor and rebooted. Now, whenever I boot, the system appears to hang at the same point - progress bar all white - but eventually (after 1+ minutes) continues to a successful boot.

Plugging the external monitor in has somehow altered the boot process, but I cannot figure out what exactly is happening. I doubt it is a BIOS problem, because the boot sequence proceeds as normal up to the point where it hangs.

Any suggestions would be gratefully received.

[secipolla]

Remove 'rghb quiet' from the 'linux' boot line ('e' in grub's screen to edit) so you can see the messages.
The point you say it sticks I think is when X is starting, so maybe the Xorg log may say something too.

[Doug Hutcheson]

Originally Posted by secipolla

Remove 'rghb quiet' from the 'linux' boot line

Excellent advice - wish I had thought of it myself. I will try it next time I can get to the machine.
Thanks for replying.
Cheers, Doug.

[Doug Hutcheson]

My original post suggested a problem with using an external monitor, but that proves to have been a red herring. Following secipolla's advice, I removed 'rhgb quiet' from the boot command and observed where the delay is occurring.

The problem appears to be with booting PostgreSQL, Apache and sendmail. I have recently started using Apache and PostgreSQL on this computer and have not rebooted since, so I fooled myself as to what was causing the boot delay.

The first slow-down is with "Started LSB: start and stop PostgreSQL server", which takes about 30 seconds to respond with OK. The second and longer pause in proceedings is with "Started LSB: start and stop Apache HTTP Server" which takes closer to a minute to respond. (These times are guesses - I have not run a stopwatch over them.) During these pauses, the system does not respond to Alt+Ctrl+F2 to display a terminal, so I am inclined to think this is a busy halt. The final issue, identified from boot.log, seems to be with "Starting sm-client: Starting LSB: start and stop sendmail", which responds quickly with "failed":

Once through these issues, the final steps and boot to graphical log-in dialog happens very quickly.

From boot.log:

Started LSB: HSQLDB-Server.
OK                                 <<<< Boot sequence runs ok to here.
Started LSB: start and stop PostgreSQL server.
OK                                 <<<< Hangs for about 30 secs before OK
Started LSB: start and stop Apache HTTP Server.
OK                                 <<<< Hangs about a minute before OK
Starting sm-client: Starting LSB: start and stop sendmail failed, see 'systemctl
status sendmail.service' for details.                         <<<< Fails quickly

For the sendmail failure, I executed the suggested sommand:

[doug@cardraeh ~]$ systemctl status sendmail.service
sendmail.service - LSB: start and stop sendmail
	  Loaded: loaded (/etc/rc.d/init.d/sendmail)
	  Active: failed since Sat, 23 Jun 2012 10:15:25 +1000; 23min ago
	 Process: 1426 ExecStop=/etc/rc.d/init.d/sendmail stop (code=exited, status=0/SUCCESS)
	 Process: 1446 ExecStart=/etc/rc.d/init.d/sendmail start (code=killed, signal=TERM)
	Main PID: 1278 (code=exited, status=0/SUCCESS)
	  CGroup: name=systemd:/system/sendmail.service
		  ├ 1542 sendmail: accepting connections
		  └ 2564 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue

I'm not sure what this sendmail.service information is telling me. At first glance, it appears to say that sendmail has not started properly, but Evolution seems quite happy to look at my local mail queue, which I would not have expected it to do if sendmail was down.

So, the questions now boil down to "how do I speed up booting PostgreSQL and Apache?" and "what is going on with sendmail?".

[DBelton]

Evolution and sendmail are 2 totally different and unrelated applications. Evolution will still work properly if sendmail isn't.

Sendmail (with just default setup) is pretty much just sending mail from applications to (usually) the root user. in case of errors.

Most systems really don't need sendmail running, unless you specifically need it and have configured it, and is pretty safe to disable.

systmectl disable sendmail.service

I have had some issues in the past where I disabled it, but something else (I think network manager) started it anyway. If you have that issue and wish it to be disabled

systemctl mask sendmail.service

[Doug Hutcheson]

Hmmmm...
I decided to bite the bullet and upgrade to FC16. Now I cannot get Apache to boot at all - see this message. Is it something I have done, or not done? I'm baffled.