Slow boot : firmware and loader

Issues related to applications and software problems and general support
username
Posts: 64
Joined: 2013/03/04 13:17:23

Slow boot : firmware and loader

Post by username » 2020/11/13 09:58:45

Hi,

I have a couple of new CentOS 8 servers running in two differents datacenters (both from the same provider OVH) and they both boot very slowly :

sudo systemd-analyze

Startup finished in 1min 11.456s (firmware) + 56.618s (loader) + 1.275s (kernel) + 4.354s (initrd) + 22.350s (userspace) = 2min 36.055s
multi-user.target reached after 22.344s in userspace

The firmware/loader part takes 2mins. Ok both servers have HDD and they are obviously slower than the ones using SSD. But still... 2 min is way more than my old CentOS 6 servers with HDD.

I have checked the HDD with smartctl and they seems to be fine.

Can it be an issue related to RAID ? An hardware issue ? Should I be worried and contact the server provider or forget about it ?

BShT
Posts: 370
Joined: 2019/10/09 12:31:40

Re: Slow boot : firmware and loader

Post by BShT » 2020/11/13 12:01:59

you can´t control someones else (cof cof cloud) computer...

i have a lot of Centos 8 servers running under ESXi and they are all OK but i have my own data center and my external data center we have an explicit contract with SSD and 1 virtual per 1 physical.

are they physical machines? are the HDDs local, iscsi or FB? SATA or SAS?

it is impossible to presume without more information

username
Posts: 64
Joined: 2013/03/04 13:17:23

Re: Slow boot : firmware and loader

Post by username » 2020/11/13 12:26:28

Sorry, I forgot to mention that they were physical dedicated servers with HDD in SATA.

ADVANCE-2 - Intel Xeon-E 2136 - 32GB DDR4 ECC 2666MHz - 2x HDD SATA 4TB Datacenter Class Soft RAID

I'm not sure about iSCI. I didn't know anything about it until this morning when I saw this error in systemd logs but maybe it is unrelated.

iscsi.service: Unit cannot be reloaded because it is inactive.: 5 Time(s)

BShT
Posts: 370
Joined: 2019/10/09 12:31:40

Re: Slow boot : firmware and loader

Post by BShT » 2020/11/13 13:14:50

try to disable services that you don´t need

iscsi.service seems to be one of them

username
Posts: 64
Joined: 2013/03/04 13:17:23

Re: Slow boot : firmware and loader

Post by username » 2020/11/13 17:09:15

Yep but that doesn't seem to be the origin of the problem here.

I will ask the hardware provider to check if the hardware is okay.

The reboot time itself isn't a big deal but I'm worried it could be a bad sign.

BShT
Posts: 370
Joined: 2019/10/09 12:31:40

Re: Slow boot : firmware and loader

Post by BShT » 2020/11/13 18:24:59

if you have console at boot, try to boot verbose

server hardware is, most of time, slow to POST but boot is normal

User avatar
TrevorH
Forum Moderator
Posts: 29915
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: Slow boot : firmware and loader

Post by TrevorH » 2020/11/13 19:34:55

Could you post the full output from systemd-analyze blame, wrapped in [code][/code] tags to preserve formatting (the last one is badly formatted and incomplete).
CentOS 6 will die in November 2020 - migrate sooner rather than later!
Info for USB installs on http://wiki.centos.org/HowTos/InstallFromUSBkey
CentOS 5 is dead, do not use it.
Full time Geek, part time moderator. Use the FAQ Luke

username
Posts: 64
Joined: 2013/03/04 13:17:23

Re: Slow boot : firmware and loader

Post by username » 2020/11/16 09:45:55

Code: Select all

$ sudo systemd-analyze blame
          7.508s docker.service
          6.160s cloud-init-local.service
          4.674s mariadb.service
          3.919s tuned.service
          3.820s php-fpm.service
          3.645s initrd-switch-root.service
          2.507s containerd.service
          2.369s kdump.service
          2.299s nginx.service
          2.254s dracut-initqueue.service
          2.132s sssd.service
          1.938s firewalld.service
          1.895s systemd-udev-settle.service
          1.464s polkit.service
          1.375s cloud-init.service
          1.239s cloud-final.service
          1.043s microcode.service
           684ms cloud-config.service
           560ms rsyslog.service
           470ms sshd.service
           467ms chronyd.service
           372ms lvm2-monitor.service
           329ms gssproxy.service
           326ms sysroot.mount
           307ms systemd-resolved.service
           289ms systemd-udevd.service
           263ms systemd-journald.service
           262ms systemd-tmpfiles-setup.service
           256ms auditd.service
           241ms rpc-statd-notify.service
           221ms rngd-wake-threshold.service
           209ms dnf-makecache.service
           171ms dev-disk-by\x2duuid-2ac8ebf6\x2d832c\x2d4afe\x2dadaf\x2dafeb3974d34b.swap
           165ms systemd-udev-trigger.service
           157ms systemd-logind.service
           145ms systemd-tmpfiles-setup-dev.service
           132ms systemd-remount-fs.service
           131ms dev-mqueue.mount
           129ms dracut-pre-pivot.service
           124ms nis-domainname.service
           114ms systemd-vconsole-setup.service
           114ms rpcbind.service
           113ms dev-disk-by\x2duuid-5e158dc0\x2dab2e\x2d4cb8\x2d99e5\x2d7af396d2af03.swap
           113ms sys-kernel-debug.mount
           108ms import-state.service
           102ms kmod-static-nodes.service
            94ms boot.mount
            90ms var-lib-nfs-rpc_pipefs.mount
            86ms unbound-anchor.service
            79ms systemd-sysctl.service
            76ms initrd-parse-etc.service
            64ms boot-efi.mount
            62ms initrd-cleanup.service
            61ms user@1001.service
            56ms dracut-cmdline.service
            47ms dev-hugepages.mount
            45ms systemd-random-seed.service
            44ms systemd-journal-flush.service
            42ms dracut-pre-trigger.service
            39ms dracut-pre-udev.service
            32ms iscsi-shutdown.service
            32ms fail2ban.service
            32ms systemd-update-utmp.service
            29ms docker.socket
            22ms sendmail.service
            20ms systemd-user-sessions.service
            15ms mdmonitor.service
            13ms systemd-fsck-root.service
            13ms sm-client.service
            12ms NetworkManager.service
            11ms dracut-shutdown.service
            10ms systemd-tmpfiles-clean.service
             9ms dracut-pre-mount.service
             7ms initrd-udevadm-cleanup-db.service
             4ms systemd-update-utmp-runlevel.service
             3ms sys-kernel-config.mount

username
Posts: 64
Joined: 2013/03/04 13:17:23

Re: Slow boot : firmware and loader

Post by username » 2020/11/17 10:48:18

There doesn't seem to be a problem with the userspace part. The long boot time comes from : 1min 11.456s (firmware) + 56.618s (loader)

But I don't know what it really refers to.

Firmware ? Could it be that CentOS 8 drivers/kernel doesn't fully support the server's hardware ?

Loader ? Could it be a problem with Grub ?

User avatar
TrevorH
Forum Moderator
Posts: 29915
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: Slow boot : firmware and loader

Post by TrevorH » 2020/11/17 11:02:02

Run dmesg -T | less and check the timestamps there to see where the delay is occurring.
CentOS 6 will die in November 2020 - migrate sooner rather than later!
Info for USB installs on http://wiki.centos.org/HowTos/InstallFromUSBkey
CentOS 5 is dead, do not use it.
Full time Geek, part time moderator. Use the FAQ Luke

Post Reply

Return to “CentOS 8 - General Support”