Opened 10 years ago

Closed 10 years ago

#4201 closed bug (fixed)

Transaction getting too big for log when checkfs corrects too many errors

Reported by: mmlr Owned by: axeld
Priority: normal Milestone: R1
Component: File Systems/BFS Version: R1/pre-alpha1
Keywords: Cc:
Blocked By: Blocking:
Has a Patch: no Platform: All

Description

On my source partition that is riddled with name mismatches, correcting them with checkfs causes a too large transaction to be generated. Just after checkfs has run through a corresponding message is printed to the syslog. When rebooting, many more such messages get printed with increasing block counts. If I don't reboot after running checkfs but use the system for a little while longer, there are sometimes panics that indicate "transaction aborted, inode xy still around!". Note that due to #4196 I don't get that far unless I disable removing invalid nodes from the BFS disk_system add-on.

KERN: transaction too large (14836 blocks, log size 2048)!
KERN: transaction too large (14837 blocks, log size 2048)!
KERN: transaction too large (14838 blocks, log size 2048)!
KERN: transaction too large (14839 blocks, log size 2048)!
KERN: transaction too large (14840 blocks, log size 2048)!
KERN: transaction too large (14841 blocks, log size 2048)!

Change History (1)

comment:1 Changed 10 years ago by axeld

Resolution: fixed
Status: newclosed

Fixed in hrev32227.

Note: See TracTickets for help on using tickets.