You are not logged in.
I have a device formatted with ext4 (no_auto_da_alloc,data=ordered).
During shutdown, when I try to force remount as read-only, the mounted filesystem. sync;echo "u" > /proc/sysrq-trigger
I see these ext4 errors after remount
76.209728 Emergency Remount R/O
76.282979 Emergency Remount Complete
77.188953 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
77.205402 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
78.220353 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
79.246658 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
80.261003 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
80.272068 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 1024 pages, ino 131067; err -30
81.300966 EXT4-fs (mmcblk0p15): ext4_da_writepages: jbd2_start: 8192 pages, ino 131067; err -30
(err 30 is -EROFS, error due to writeback to read-only filesystem).
As part of remount, I think linux will writeback all drity datas. Should no writeback occur after remount R/O. Not sure why I see these errors.
Any help is appreciated.
Offline
I would expect that sync flushes all dirty data to disk. jbd2_start() is related to journal though, maybe the filesystem tries to replay the journal log or something like this?
A few things you can do to understand the issue better. Find what inode the process try to write, use 'find -inum 131067' does the filename rings the bell?
If you are still enthusiastic about debugging the issue then the next step is to enable kernel traces to see 1) what exactly process writes the data 2) where in kernel it is stuck.
https://www.kernel.org/doc/Documentatio … events.txt
https://www.kernel.org/doc/Documentatio … ftrace.txt
Read it before posting http://www.catb.org/esr/faqs/smart-questions.html
Ruby gems repository done right https://bbs.archlinux.org/viewtopic.php?id=182729
Fast initramfs generator with security in mind https://wiki.archlinux.org/index.php/Booster
Offline
BTW How
echo "u" > /proc/sysrq-trigger
differs from
mount -a -o remount,ro?
Read it before posting http://www.catb.org/esr/faqs/smart-questions.html
Ruby gems repository done right https://bbs.archlinux.org/viewtopic.php?id=182729
Fast initramfs generator with security in mind https://wiki.archlinux.org/index.php/Booster
Offline
Find what inode the process try to write, use 'find -inum 131067' does the filename rings the bell?
Yes, I decide to find out the file with inode 131067, but it's sporadic, need the next time happened.
From the log, seems like if ext4_da_writepages failled, these pages will be re-writed later?
Offline