[SOLVED] Too Long Boot Time [15]

[dedman]

Hi.
Im a novice and migrated from ubuntu to fedora a few days back and now lovin fedora.
Also, recently i installed a few packages that were mainly programming libraries.

Earlier my boot time was well under 40 seconds. However now it has strangely risen to above 2.5 minutes.

Can someone please help me figure out , what has gone wrong.

I have uploaded the bootchart image : http://img705.imageshack.us/img705/5686/bootchartj.png

Also the the file BOOT.LOG is attached.

The Output from dmesg also has been attached.

Please forgive me for my lack of knowledge on this.

[dedman]

a gentle bounce!

[glennzo]

For some reason I can't view the boot chart graphic on my laptop. I looked at boot.log and dmesg.txt. There's probably some hints there for those more experienced than I, although looking at dmesg.txt the system may be hanging here

[ 27.864012] ppdev: user-space parallel port driver
[ 160.566133] hda-intel: IRQ timing workaround is activated for card #1. Suggest a bigger bdl_pos_adj.

here

[ 168.951529] bootchart-collector pid: 1340 unmounted proc / clean exit
[ 301.548960] fuse init (API version 7.16)

or here

[ 301.600624] SELinux: initialized (dev fuse, type fuse), uses genfs_contexts
[ 818.597428] usb 2-1.3: new high speed USB device number 3 using ehci_hcd

Or it could be hanging at all three points. My first suggestion would be to boot using text mode. Do that by editing /boot/grub/grub.conf and removing the "rhgb quiet" and replacing it with a space and "vga=788", minus the quotes. The system will still boot to the GUI but you will be able to see all the text flying by and will be able to see where it is hanging. Might be helpful.

[weitjong]

Could you post the output from this command.

$ systemd-analyze blame

It show you which service(s) should take the blame of your long boot time.

[cneil]

Originally Posted by dedman

Hi.
Im a novice and migrated from ubuntu to fedora a few days back and now lovin fedora.
Also, recently i installed a few packages that were mainly programming libraries.

Earlier my boot time was well under 40 seconds. However now it has strangely risen to above 2.5 minutes.

Can someone please help me figure out , what has gone wrong.

I have uploaded the bootchart image : http://img705.imageshack.us/img705/5686/bootchartj.png

Also the the file BOOT.LOG is attached.

The Output from dmesg also has been attached.

Please forgive me for my lack of knowledge on this.

In a terminal type in 'top' and look at the result. It will tell you what is running and what is not.

[dedman]

Thanks a lot for looking into my problem.

I have looked in to the systemd-analyze blame command's output.

It seems SENDMAIL.SERVICE is taking too long.

Here is the output :

121005ms sendmail.service
  5328ms udev-settle.service
  5127ms systemd-vconsole-setup.service
  4094ms fedora-readonly.service
  3375ms bootchart.service
  3301ms remount-rootfs.service
  2815ms media.mount
  2752ms systemd-remount-api-vfs.service
  2737ms cups.service
  2462ms udev.service
  2061ms fedora-sysinit-hack.service
  1594ms systemd-sysctl.service
  1593ms hwclock-load.service
  1231ms udev-trigger.service
   865ms bluetooth.service
   862ms rsyslog.service
   817ms avahi-daemon.service
   773ms abrtd.service
   678ms NetworkManager.service
   492ms fedora-storage-init.service
   422ms fedora-sysinit-unhack.service
   387ms lvm2-monitor.service
   376ms console-kit-log-system-start.service
   323ms systemd-tmpfiles-setup.service
   304ms dbus.service
   289ms livesys.service
   259ms irqbalance.service
   241ms mcelog.service
   236ms fedora-storage-init-late.service
   235ms ip6tables.service
   226ms iptables.service
   221ms cpuspeed.service
   181ms mdmonitor.service
   179ms auditd.service
   172ms netfs.service
   171ms systemd-readahead-collect.service
   171ms iscsi.service
   167ms lldpad.service
   159ms abrt-ccpp.service
   155ms iscsid.service
   135ms portreserve.service
   132ms sandbox.service
   131ms systemd-readahead-replay.service
   127ms console-kit-daemon.service
   118ms abrt-oops.service
   105ms fedora-wait-storage.service
    88ms smolt.service
    84ms rc-local.service
    82ms fedora-autoswap.service
    43ms rtkit-daemon.service
    39ms livesys-late.service
    30ms accounts-daemon.service
    19ms systemd-user-sessions.service

[jpollard]

Sendmail doesn't stop. It just starts and continues.

Most workstations don't need sendmail, though it is useful to receive local messages from cron.

One of the issues with sendmail is that it needs the TCP service to be funtctional, and if the network is not, then it runs into difficulty looking up names and identifying itself. This can cause a long delay in certain startup modes.

[dedman]

Also, i tried booting the older version of the kernel, but no improvement in boot time. However i got this at the screen.

Started SYSV : Late init script for Live Image

http://img9.imageshack.us/img9/3016/dsc00014yf.jpg

---------- Post added at 12:58 AM ---------- Previous post was at 12:56 AM ----------

Originally Posted by jpollard

Sendmail doesn't stop. It just starts and continues.

Most workstations don't need sendmail, though it is useful to receive local messages from cron.

One of the issues with sendmail is that it needs the TCP service to be funtctional, and if the network is not, then it runs into difficulty looking up names and identifying itself. This can cause a long delay in certain startup modes.

I do not have a working internet connection during boot, since i connect using a USB Modem. Could it cause the issue, you are talking about ??
What do you suggest in such a case ?

[jpollard]

You can try disabling the service.

Since sendmail will try to resolve it's own name, what you CAN do is put the local host name in /etc/hosts as 127.0.0.1. This can sometimes resolve DNS timeout issues.

[dedman]

Originally Posted by jpollard

You can try disabling the service.

Since sendmail will try to resolve it's own name, what you CAN do is put the local host name in /etc/hosts as 127.0.0.1. This can sometimes resolve DNS timeout issues.

Well, this is how the file looks like. The 127.0.0.1 has already been added twice.

Here is the screenshot : http://img18.imageshack.us/img18/8560/etcn.png

Is the ::1 line causing the problem, or adding the address 127.0.0.1 twice ??

[jpollard]

Neither should be causing a problem.

I suspect the problem is elsewhere, but due to the parallel problem of systemd, it will be hard to find.

You can try disabling each active service until you find the one that unlocks things.

[weitjong]

I had the exactly the same problem with the sendmail service a few weeks ago. Look at the sendmail log file in the "/var/log/maillog". For my case back then, I was seeing some time out entries due to the sendmail service not able to reach my host machine by its hostname. If that is also your problem then the fix is easy. Just add your hostname to the end of localhost entries in "/etc/hosts" file.

For example, this is what I have. Replace "igloo" with your own hostname.

[root@igloo etc]# grep igloo hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4 igloo
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6 igloo

[dedman]

Originally Posted by jpollard

Neither should be causing a problem.

I suspect the problem is elsewhere, but due to the parallel problem of systemd, it will be hard to find.

You can try disabling each active service until you find the one that unlocks things.

Originally Posted by weitjong

I had the exactly the same problem with the sendmail service a few weeks ago. Look at the sendmail log file in the "/var/log/maillog". For my case back then, I was seeing some time out entries due to the sendmail service not able to reach my host machine by its hostname. If that is also your problem then the fix is easy. Just add your hostname to the end of localhost entries in "/etc/hosts" file.

For example, this is what I have. Replace "igloo" with your own hostname.

[root@igloo etc]# grep igloo hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4 igloo
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6 igloo

Thanks a lot for the help. I checked out the file. and added the hostname. I also removed 127.0.0.1 that was added redundantly by my XAMPP installation. This did the trick. Thanks guys for the help.

[maeniel]

Thanks for this, solved my issue as well.

[Mariusz W]

In my case sendmail.service takes 7.5 longer to initialize (61 seconds) than the next slowest service, udev-settle.service, which takes 8 seconds.

The list of services is long: on my default Fedora 16 installation exactly 54 are initialized at boot time. If to add up all the initialization times according to systemd-analyze blame, then one gets 156 seconds.

The actual boot time however is significatly faster, around 20-25 seconds.