CentOS: very slow system boot / sometimes hangs entirely

General support questions
socrates
Posts: 13
Joined: 2017/10/24 11:28:06

CentOS: very slow system boot / sometimes hangs entirely

Postby socrates » 2018/01/19 13:44:40

Hi,

I have CentOS installed on a GigaByte N3150-D3V Board.

For some reason I can't figure out, the system boots *very* slowly and sometimes even hangs. Click here for a demonstration video of the slow boot.

My gut tells me it has something to do with the video framebuffer, so I tried to disable it by various setting (blacklist i915 in /etc/modprobe.d/blacklist.conf), following this Guide or this one for ubuntu.

But nothing helps neither as far as the slow boot is concerned nor as far as disabling framebuffer and the pinguins.

I ran a systemd-analyze plot > heBacula-boot-analysis.svg and apparently kernel (with 33 sec) and initrd (47 sec) take the most time to complete.

There is also the output of dmesg which I examined but there's also nothing that looks suspicious to me.


Anyone got any pointers as to what the cause of the problem could be?

Any help would be greatly appreciated!

Thanks in advance,

socrates

User avatar
TrevorH
Forum Moderator
Posts: 21773
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: CentOS: very slow system boot / sometimes hangs entirely

Postby TrevorH » 2018/01/19 13:56:50

Almost all the big gaps in timestamps in your dmesg output seem to be to do with USB devices.

Your svg file tells me nothing. What's the text output from systemd-analyze blame instead?
CentOS 5 died in March 2017 - migrate NOW!
Full time Geek, part time moderator. Use the FAQ Luke

socrates
Posts: 13
Joined: 2017/10/24 11:28:06

Re: CentOS: very slow system boot / sometimes hangs entirely

Postby socrates » 2018/01/19 15:13:27

TrevorH wrote: Your svg file tells me nothing.


Well, the only thing I found interesting in the SVG file was in the upper left corner:

Code: Select all

Startup finished in 16.682s (firmwre) + 2.831s (loader) + 33.105s (kernel) + 47.164s (initrd) + 16.892s (userspace) = 1 min 56.677s


The rest is pretty pointless, I agree.

I didn't include the ouptut from systemd-analyze blame because the "heavy" part is not included here:


TrevorH wrote: What's the text output from systemd-analyze blame instead?


Code: Select all

9.289s systemd-vconsole-setup.service
          5.676s NetworkManager-wait-online.service
          3.669s mariadb.service
          3.598s kdump.service
          1.617s postfix.service
          1.424s network.service
          1.104s lvm2-monitor.service
          1.103s dev-mapper-vg_os\x2droot.device
           933ms firewalld.service
           672ms lvm2-pvscan@9:124.service
          9.289s systemd-vconsole-setup.service
          5.676s NetworkManager-wait-online.service
          3.669s mariadb.service
          3.598s kdump.service
          1.617s postfix.service
          1.424s network.service
          1.104s lvm2-monitor.service
          1.103s dev-mapper-vg_os\x2droot.device
           933ms firewalld.service
           672ms lvm2-pvscan@9:124.service
           619ms tuned.service
           494ms srv.mount
           429ms boot.mount
           373ms lvm2-pvscan@9:127.service
           319ms var.mount
           304ms auditd.service
           258ms systemd-logind.service
           222ms NetworkManager.service
           217ms tmp.mount
           213ms var-log.mount
           196ms systemd-journald.service
           179ms proc-fs-nfsd.mount
           178ms kmod-static-nodes.service
           175ms dev-hugepages.mount
           161ms systemd-tmpfiles-clean.service
           143ms polkit.service
           141ms systemd-udev-trigger.service
           140ms dev-mqueue.mount
           132ms sshd.service
           126ms systemd-tmpfiles-setup.service
           125ms rhel-import-state.service
           121ms var-log-audit.mount
           110ms home.mount
            99ms systemd-tmpfiles-setup-dev.service
            99ms systemd-udevd.service
            98ms rhel-readonly.service
            95ms yum-cron.service
            94ms mdmonitor.service
            90ms microcode.service
            88ms ntpd.service
            85ms gssproxy.service
            85ms rhel-dmesg.service
            83ms rpc-statd-notify.service
            83ms bacula-sd.service
            82ms systemd-readahead-done.service
            77ms dev-disk-by\x2duuid-82a895cc\x2d2acb\x2d45a3\x2d90df\x2d7be08a242de0.swap
            67ms bacula-fd.service
            59ms plymouth-start.service
            58ms bacula-dir.service
            57ms boot-efi_disk_SN_WCC4N7JCCFY9.mount
            54ms systemd-readahead-collect.service
            53ms systemd-readahead-replay.service
            53ms rsyslog.service
            50ms boot-efi_disk_SN_WCC4N7JCCZC9.mount
            50ms boot-efi.mount
            47ms sys-kernel-debug.mount
            47ms systemd-sysctl.service
            47ms boot-efi_disk_SN_WCC4N2ATJA1C.mount
            44ms systemd-remount-fs.service



TrevorH wrote:Almost all the big gaps in timestamps in your dmesg output seem to be to do with USB devices.


Right. The kernel even claims to have found a Logitech mouse:

Code: Select all

[   74.156209] input: PS/2 Logitech Mouse as /devices/platform/i8042/serio1/input/input5


But the only thing attached to the server are network, monitor and (cherry) keyboard.

hunter86_bg
Posts: 905
Joined: 2015/02/17 15:14:33
Location: Bulgaria
Contact:

Re: CentOS: very slow system boot / sometimes hangs entirely

Postby hunter86_bg » 2018/01/21 20:22:25

Can you try with the rescue kernel and check the time needed to boot ?
Did you notice this behaviour in earlier kernels or it is observed with the latest one ?