Hide Forgot
Description of problem: After patching one of my servers that uses ext4 filesystems with the newest kernel (2.6.18-238.12.1.el5) I began getting the following messages in the syslog: kernel: JBD: barrier-based sync failed on <disk> - disabling barriers I rebooted using the previous kernel and the messages did not appear again. This is occurring on my SAN disks using multipath devices. I don't know if it is isolated to them or not. Version-Release number of selected component (if applicable): RHEL 5.5, kernel 2.6.18-238.12.1.el5 How reproducible: Happens every time during boot up. Steps to Reproduce: 1. Install kernel 2.6.18-238.12.1.el5 on system with ext4 filesystems using multipath devices 2. Reboot 3. Actual results: Expected results: Additional info:
Correction: This is not SAN storage or multipath devices. It is local (single path) disk devices.
This is an expected (informational) message when the storage does not support barriers. It should work on local storage (S-ATA and SAS drives), can you run "hdparm -I /dev/sda" and add that in. Please open a support ticket through RH officially. RH bugzilla is not meant to be a customer facing support tool since that cuts our field support out of the loop. Thanks!
After doing some further researching I discovered the devices it is complaining about aren't disk devices that the system is using at all because checking /proc/mounts shows that all EXT4 filesystems have barriers enabled (barrier=1). So the new kernel is checking for barrier support on devices that the previous kernel was not. The devices reported were in the format dm-#-#, such as dm-0-8. I'm assuming this is some type of dm-multipath device but it doesn't refer to any of my disk devices. Anyway, it isn't affecting us so feel free to close this ticket. Thanks.
Thanks Darren - in RHEL6, barriers will work for dm and lvm smoothly as well.