1

I should note that this is a dual boot system with Windows 10.

Here's my latest systemd-analyze

Startup finished in 20.022s (firmware) + 2.835s (loader) + 2.215s (kernel) + 36.136s (userspace) = 1min 1.211s

Here's my latest systemd-analyze blame

         56.166s upower.service
         35.175s lightdm.service
         35.157s plymouth-quit-wait.service
           996ms systemd-rfkill.service
           637ms dev-sdb5.device
           102ms ModemManager.service
            98ms accounts-daemon.service
            94ms networking.service
            93ms systemd-udevd.service
            89ms NetworkManager.service
            76ms grub-common.service
            76ms thermald.service
            75ms apport.service
            68ms systemd-journald.service
            50ms systemd-udev-trigger.service
            46ms keyboard-setup.service
            45ms systemd-logind.service
            45ms avahi-daemon.service
            38ms irqbalance.service
            37ms apparmor.service
            36ms systemd-tmpfiles-clean.service
            32ms console-setup.service
            31ms ondemand.service
            28ms speech-dispatcher.service
            27ms systemd-fsck@dev-disk-by\x2duuid-8655\x2d1473.service
            26ms systemd-modules-load.service
            23ms systemd-tmpfiles-setup-dev.service
            22ms udisks2.service
            22ms plymouth-start.service
            22ms gpu-manager.service
            18ms alsa-restore.service
            15ms pppd-dns.service
            15ms systemd-user-sessions.service
            12ms rsyslog.service
            11ms plymouth-read-write.service
            10ms systemd-update-utmp.service
            10ms polkitd.service
             9ms colord.service
             9ms dev-mqueue.mount
             9ms user@1000.service
             9ms bluetooth.service
             8ms systemd-timesyncd.service
             8ms wpa_supplicant.service
             7ms systemd-sysctl.service
             7ms systemd-journal-flush.service
             7ms ufw.service
             6ms sys-kernel-debug.mount
             6ms ureadahead-stop.service
             5ms systemd-tmpfiles-setup.service
             5ms kmod-static-nodes.service
             4ms snapd.autoimport.service
             3ms systemd-random-seed.service
             3ms systemd-update-utmp-runlevel.service
             3ms dev-hugepages.mount
             3ms dev-sdb6.swap
             3ms systemd-remount-fs.service
             2ms resolvconf.service
             2ms dns-clean.service
             2ms boot-efi.mount
             1ms rtkit-daemon.service
             1ms sys-fs-fuse-connections.mount
             1ms setvtrgb.service
             1ms rc-local.service
           351us snapd.socket

The times don't really add up and there's always different services taking a long time. One other boot has ModemManager.service taking 50 seconds. Another boot had a different process taking 1 min 30.

I have uninstalled and reinstalled multiple times and I can't figure it out. Help is appreciated and beware I am a new user.

Jason
  • 11
  • 2
  • A wide variance in delays may be related to a failing storage device. Have you checked the content of /var/log/syslog for clues? Checked smart status? – Elder Geek Jul 25 '17 at 12:11
  • What are some telltale signs I should look for in syslog? I ran the SMART tests on my SSD and it comes back fine. It's a relatively new SSD, just under a year old. Crucial MX300 750GB. – Jason Jul 25 '17 at 13:11
  • https://askubuntu.com/questions/804918/how-to-filter-out-errors-in-ubunty-system-log-files – Elder Geek Jul 25 '17 at 13:48
  • I see this Jul 25 13:17:10 Jason-Ubuntu kernel: [ 69.302347] ata8.00: irq_stat 0x08000000, interface fatal error, I see this Jul 25 12:27:13 Jason-Ubuntu kernel: [ 2.496598] EXT4-fs (sdb5): re-mounted. Opts: errors=remount-ro. I see a block of these: Jul 25 12:27:44 Jason-Ubuntu kernel: [ 34.276618] blk_update_request: I/O error, dev sdb, sector 1413962872 and a lot of stuff to do with gnome and my AMD gpu drivers. – Jason Jul 25 '17 at 13:59
  • 1
    Ideally you should [edit] this information into your question. as comments can ge deleted for numerous reasons. Also the output of sudo lshw -C disk and sudo fdisk -l as based on your comment there's clearly an issue with /dev/sdb (or the cable connecting it) That's the drive you need to be testing. – Elder Geek Jul 25 '17 at 16:20
  • I managed to figure it out. NCQ was an issue, disabling it fixed it. Seemed to be common with my type of SSD. – Jason Jul 28 '17 at 23:55
  • if you resolved your issue, why not [edit] your question to include the type of SSD (comments can get deleted for numerous reasons and are unlikely to appear in search engine results) that you have that this is common with and write an answer so that others with the same issue can benefit from your experience? – Elder Geek Jul 29 '17 at 13:28

1 Answers1

0

Your /dev/sdb is getting I/O errors. That's not good for a new SSD.

First, review the SMART data and run the short test. Open the Disks app, select the SSD in the left pane, go to the "hamburger" icon, select SMART Data & Tests. Review the data, and run the short test. Warning... do NOT run a badblock test on a SSD.

Second, go to the Crucial web site at http://www.crucial.com/usa/en/support-ssd-firmware, and download the latest firmware dated 5/16/2017. The new firmware better handles exactly the errors that you're getting. FIRST backup your important data, THEN install the new firmware.

Continue to monitor /var/log/syslog for sdb errors...

In terminal...

grep -i sdb /var/log/syslog*

If they continue, log a warranty ticket with Crucial.

Update #1:

Problem was ultimately solved with kernel parameter libata.force=noncq.

heynnema
  • 70,711
  • I figured it out, it's a problem with NCQ. I disabled that and it boots fine. On Windows my SSD is fine, other people had similar issues with the same SSD on Linux. – Jason Jul 28 '17 at 23:56
  • How did you disable NCQ? Did you update the firmware in the SSD? If not, it still needs to be done... as it handles errors like yours, and maybe you won't need to disable NCQ. Report back. – heynnema Jul 29 '17 at 00:00
  • I disabled NCQ with a kernal boot parameter. libata.force=noncq. I updated firmware through windows beforehand and it didn't solve the issue. – Jason Jul 30 '17 at 00:40