Bug 111841 - bogus errors 2001, no inode item, link count wrong, could be race
Summary: bogus errors 2001, no inode item, link count wrong, could be race
Status: NEW
Alias: None
Product: File System
Classification: Unclassified
Component: btrfs (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Josef Bacik
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-02-03 21:15 UTC by Chris Murphy
Modified: 2016-03-20 09:53 UTC (History)
1 user (show)

See Also:
Kernel Version: 4.4.0-1.fc24.x86_64
Subsystem:
Regression: No
Bisected commit-id:

Attachments
screeshot of problem (1.13 MB, image/jpeg)
2016-02-03 21:15 UTC, Chris Murphy 		Details
Add an attachment (proposed patch, testcase, etc.)

Description Chris Murphy 2016-02-03 21:15:37 UTC 
Created attachment 202901 [details]
screeshot of problem

btrfs-progs 4.4, but has been happening for a while, at least since btrfs-progs 4.0

1. Boot to dracut shell using rd.break=pre-mount
2. btrfs check <dev>
## hundreds of messages (see screen shot attached) such as:

root 402 inode 151731 errors 2001, no inode item, link count wrong
        unresolved ref dir 16314 index 341 namelen 13 name os2empath.py filetype 1 errors 4, no inode ref

inode # increments by 1 with each error, the unresolved ref dir appears to be always be the same.

3. Cancel with command-C
4. Rerun the command again
## no errors



Additional Info:
[    0.830362] scsi 0:0:0:0: Direct-Access     ATA      Samsung SSD 840  DB6Q PQ: 0 ANSI: 5
[    0.831331] ata3.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
[    0.831344] ata3.00: configured for UDMA/100
[    0.833909] usb 1-1: new high-speed USB device number 2 using ehci-pci
[    0.836527] sd 0:0:0:0: Attached scsi generic sg0 type 0
[    0.836577] sd 0:0:0:0: [sda] 488397168 512-byte logical blocks: (250 GB/233 GiB)
[    0.837185] sd 0:0:0:0: [sda] Write Protect is off
[    0.837197] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[    0.837451] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Note You need to log in before you can comment on or make changes to this bug.