Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FS#3140 - sysupgrade spews lost of errors on x86_64 #8298

Open
openwrt-bot opened this issue Jun 1, 2020 · 5 comments
Open

FS#3140 - sysupgrade spews lost of errors on x86_64 #8298

openwrt-bot opened this issue Jun 1, 2020 · 5 comments
Labels

Comments

@openwrt-bot
Copy link

pprindeville:

This is on an x86_64 build of master from last Monday (circa 6e8b689).

The exact hardware is a Supermicro SYS-5018D-FN8T, 32GB of DRAM, 512GB of NVMe.

I built a new master today (8d2c031) and scp'd it onto /tmp/ of my router. Then did a "sysupgrade /tmp/..." of that:

root@OpenWrt:/# ifdown wan
root@OpenWrt:/#
root@OpenWrt:/#
root@OpenWrt:/# sysupgrade /tmp/openwrt-r13360+64-4661b05390-x86-64-generic-squashfs-combined-efi.img
Image metadata not found
Reading partition table from bootdisk...
cat: write error: Broken pipe
Reading partition table from image...
Saving config files...
Commencing upgrade. Closing all shell sessions.
Watchdog handover: fd=7

  • watchdog -
    killall: telnetd: no process killed
    killall: dropbear: no process kiSending TERM to remaining processes ... ntpd sh snmpd dhcpd ipset-dns collectd lldpd lldpd smartd named sleep ubusd askfirst urngd netifd rngd crond lighttpd sshd syslog-ng
    Sending KILL to remaining processes ...
    Switching to ramdisk...
    Performing system upgrade...
    Reading partition table from bootdisk...
    dd: warning: partial read (8192 bytes); suggest iflag=fullblock
    0+63 records in
    0+63 records out
    1101824 bytes (1.1 MB, 1.1 MiB) copied, 0.00492347 s, 224 MB/s
    Reading partition table from image...
    Writing image to /dev/nvme0n1p1...
    cat: write error: Broken pipe
    40959+0 records in
    19+1 records out
    20971008 bytes (21 MB, 20 MiB) copied, 0.0963313 s, 218 MB/s
    Writing image to /dev/nvme0n1p2...
    462847+0 records in
    225+1 records out
    236977664 bytes (237 MB, 226 MiB) copied, 0.810288 s, 292 MB/s
    Writing new UUID to /dev/nvme0n1...
    4+0 records in
    4+0 records out
    cat: write error: Broken pipe
    4 bytes copied, 0.00439713 s, 0.9 k[ 2308.927159] F2FS-fs (nvme0n1p1): Magic Mismatch, valid(0xf2f52010) - read(0x0)
    B/s
    [ 2308.945667] F2FS-fs (nvme0n1p1): Can't find valid F2FS filesystem in 1th superblock
    [ 2308.963141] F2FS-fs (nvme0n1p1): Magic Mismatch, valid(0xf2f52010) - read(0x6020601)
    [ 2308.980487] F2FS-fs (nvme0n1p1): Can't find valid F2FS filesystem in 2th superblock
    Upgrading bootloader on /dev/nvme0n1...
    touch: /tmp/boot/grub/upgraded: No such file or directory
    Upgrade completed
    Rebooting system...
    umount: can't unmount /dev: Resource busy
    umount: can't unmount /tmp: Resource busy
    [ 2310.852582] ACPI Warning: _SB.PCI0.BR2C._PRT: Return Package has no elements (empty) (20190816/nsprepkg-96)
    [ 2310.881404] ACPI Warning: _SB.PCI0.BR2C._PRT: Return Package has no elements (empty) (20190816/nsprepkg-96)
    [ 2311.080235] reboot: Restarting system
    [ 2311.093088] reboot: machine restart

Note that on reboot, dmesg says:

...
May 31 16:43:17 OpenWrt kernel: [ 3.721825] Copyright (c) 1999-2008 LSI Corporation
May 31 16:43:17 OpenWrt kernel: [ 3.724539] nvme nvme0: 7/0/0 default/read/poll queues
May 31 16:43:17 OpenWrt kernel: [ 3.732035] Fusion MPT SPI Host driver 3.04.20
May 31 16:43:17 OpenWrt kernel: [ 3.752418] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
May 31 16:43:17 OpenWrt kernel: [ 3.753680] GPT:Primary header thinks Alt. header is not at the end of the disk.
May 31 16:43:17 OpenWrt kernel: [ 3.764418] ehci-pci: EHCI PCI platform driver
May 31 16:43:17 OpenWrt kernel: [ 3.777281] GPT:504894 != 1000215215
May 31 16:43:17 OpenWrt kernel: [ 3.787218] ehci-pci 0000:00:1a.0: EHCI Host Controller
May 31 16:43:17 OpenWrt kernel: [ 3.795949] GPT:Alternate GPT header not at the end of the disk.
May 31 16:43:17 OpenWrt kernel: [ 3.795950] GPT:504894 != 1000215215
May 31 16:43:17 OpenWrt kernel: [ 3.806441] ehci-pci 0000:00:1a.0: new USB bus registered, assigned bus number 1
May 31 16:43:17 OpenWrt kernel: [ 3.817622] GPT: Use GNU Parted to correct GPT errors.
May 31 16:43:17 OpenWrt kernel: [ 3.817629] nvme0n1: p1 p2 p128
...

but I don't know if that's related or not.

@openwrt-bot
Copy link
Author

pprindeville:

Can someone please follow up on this?

@openwrt-bot
Copy link
Author

pprindeville:

Also, can we raise the priority? If the partition table is being written in an inconsistent way, or the image isn't being completely written to disk, that's potentially fairly serious.

@openwrt-bot
Copy link
Author

pprindeville:

Still seeing a lot of spew, even with building from HEAD on 2020-10-02.

@openwrt-bot
Copy link
Author

yousong:

I just checked with the code. The errors and warnings are disturbing but should be mostly fine with regard to integrity and functionality of the system.

The following was caused by command "cat xx | dd count=x ..". It can also be zcat if the image is suffixed with .gz

cat: write error: Broken pipe

The following lines were likely caused by "mount /dev/xx /tmp/boot". It's mount probe function debug output. With combined-efi image, the boot partition will be formated as fat (or vfat). The mount will succeed.

[ 2308.945667] F2FS-fs (nvme0n1p1): Can't find valid F2FS filesystem in 1th superblock
[ 2308.963141] F2FS-fs (nvme0n1p1): Magic Mismatch, valid(0xf2f52010) - read(0x6020601)
[ 2308.980487] F2FS-fs (nvme0n1p1): Can't find valid F2FS filesystem in 2th superblock

The following is a minor flaw that the filepath is wrong, but it should not cause bad consequence.

touch: /tmp/boot/grub/upgraded: No such file or directory

The following one I cannot reproduce. But likely it's also expected as ptgen the tool we compile and use for partition table generation does not write backup (alt.) header at end of the disk.

May 31 16:43:17 OpenWrt kernel: [ 3.753680] GPT:Primary header thinks Alt. header is not at the end of the disk.

@openwrt-bot
Copy link
Author

yousong:

Pushed a fix for the "upgraded" mark, http://git.openwrt.org/bd0b8200687c69df543a7e3ef043baa39d7784be

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant