[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Interesting interaction between journal recovery and slow boots



Brian Murrell writes:
> For some time now I have been puzzled as to why certain portions
> of my system boot were quite slow -- but only after journal
> recoveries.
> 
> After much investigation (strace and "set -x" in bash scripts) and
> many times hitting the power switch) I finally figured it out.  The
> portmapper and nfslock scripts are the first ones started after
> syslogd and klogd are started.  During the journal recovery there
> are copious quantities of messages such as:
> 
> Mar 20 12:27:20 brian-laptop fsck: JFS DEBUG: (recovery.c, 366): do_one_pass:  
> Mar 20 12:27:20 brian-laptop fsck: Scanning for sequence ID 484035 at 9590/10000 
> Mar 20 12:27:20 brian-laptop fsck: JFS DEBUG: (recovery.c, 366): do_one_pass: Scanning for sequence ID 484036 at 9591/10000 
> Mar 20 12:27:20 brian-laptop fsck: JFS DEBUG: (recovery.c, 366): do_one_pass: Scanning for sequence ID 484036 at 9593/10000 

This is purely a user-space issue, coming from e2fsck it appears.  I have
no such copious logs on my system.  Maybe you should undefine the JFS_DEBUG
in e2fsprogs/include/linux/jfs.h?  I seem to recall that there was a problem
with jfs.h, where it enabled JFS_DEBUG whether in kernel mode or not.  There
is a configure option "--enable-jfs-debug" which should set JFS_DEBUG for us,
so no need to always enable it in jfs.h.

Cheers, Andreas
-- 
Andreas Dilger  \ "If a man ate a pound of pasta and a pound of antipasto,
                 \  would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/               -- Dogbert





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]