Fixes an erroneous free() that can crash lessfs upon startup when the tiger hash is selected. Changes mklessfs so that it supports automatic directory creation and database overwrites. mklessfs now has improved error reporting. And some general code cleanups.
-
Archives
- January 2015
- August 2014
- February 2014
- December 2013
- May 2013
- March 2013
- January 2013
- December 2012
- July 2012
- June 2012
- April 2012
- January 2012
- October 2011
- September 2011
- August 2011
- June 2011
- May 2011
- April 2011
- March 2011
- February 2011
- January 2011
- December 2010
- October 2010
- September 2010
- August 2010
- July 2010
- June 2010
- April 2010
- March 2010
- February 2010
- January 2010
- December 2009
- November 2009
- October 2009
- September 2009
- August 2009
- June 2009
- March 2009
-
Meta
Hi Mark, nice to read that
gonna update to the last version as soon as i’m back 
BTW, have a nice Christmas,
Alex from the Alps snow…
Anyone have installation instructions for Ubuntu 9.04? I’m interested in checking lessfs out.. do I have to compile tokyocabinet and fuse from source? Can I keep them separate from the current fuse system?
Thanks
I’me trying to install lessfs on centos 5.4 64bit by compiling from source tokyocabinet and fuse but when I lauch ./configure (with options) it can find both fuse and tokyocabinet.
Have you got ideas? I’me compiling fuse and tokyocabinet with:
1) ./configure
2) make clean
3) make install
No problem on Ubuntu 9.10 (it’s quite slowly…) with same parameters.
I have ready made RPMS for a CentOS 5.4 64-bit, with fuse and tokyocabinet.
There are available at http://download.netservant.fi/lessfs/
Its not a YUM repository format but just download needed RPM-files and make local RPM install.
Hi,
yum repo with lessfs is here:
http://fs12.vsb.cz/hrb33/el5/hrb/stable/i386/repoview/letter_f.group.html
http://fs12.vsb.cz/hrb33/el5/hrb/stable/x86_64/repoview/letter_f.group.html
Don’t forget than on Centos the kernel does not support >4096 page size.
David Hrbáč
Thx a lot for your work!
Nice that you provide CentOS RPMS.
Also nice that the lessfs-devel also provice specfiles and src.rpms
you made my day
Ok, I’ve read this file and I’ve tried to apply different “mount” opptions for lessfs on ubuntu.
Another question: it have sense to try lessfs on centos64 or not? (in front of Ubuntu 64bits).
Many thanks!
P.S.: for backup what hardware do you have? (number of CPUs/cores, ram, disk for system, etc….)
I’ve installed lessfs on ubuntu 9.10 64bit but I need to optimize it to backup many little files (order of KB).
Can you give me suggestions?
Here’s my configuration:
DEBUG = 1
HASHLEN = 24
BLOCKDATA_PATH=/lessfs/dta
BLOCKDATA_BS=10485760
BLOCKUSAGE_PATH=/lessfs_cfg/mta
BLOCKUSAGE_BS=10485760
DIRENT_PATH=/lessfs_cfg/mta
DIRENT_BS=10485760
FILEBLOCK_PATH=/lessfs_cfg/mta
FILEBLOCK_BS=10485760
META_PATH=/lessfs_cfg/mta
META_BS=10485760
HARDLINK_PATH=/lessfs_cfg/mta
HARDLINK_BS=10485760
SYMLINK_PATH=/lessfs_cfg/mta
SYMLINK_BS=10485760
FREELIST_PATH=/lessfs_cfg/mta
FREELIST_BS=10485760
CACHESIZE=640
COMMIT_INTERVAL=300
LISTEN_IP=127.0.0.1
LISTEN_PORT=100
MAX_THREADS=4
DYNAMIC_DEFRAGMENTATION=off
COREDUMPSIZE=256000000
SYNC_RELAX=0
ENCRYPT_DATA=off
ENCRYPT_META=on
lessfs /etc/lessfs.cfg /backup -o negative_timeout=0,entry_timeout=0,attr_timeout=0,use_ino,readdir_ino,default_permissions,allow_other,max_read=65536,max_write=65536
For big files (order of 60GB for each one) I’ve no problem.
Thanks for your help.
With small files it makes sense to use a smaller block size. For example 16k or 32k.
You might also want to change sync_relax=1 because this speeds up opening and closing small files. I don’t know what hardware you are using but MAX_THREADS=4 should only be used when you have more the 4 cpu cores in the system.
If you can try adding more memory to the cache. CACHESIZE=xxxx
You are using the tokyocabinet backend and sometimes more memory for caching speeds things up.
Mark.
CPU cores ar not problem: I’ve 2 CPU with 2core for each one.
I’ve many files less than 10KB and to backup 45GB of Directories it use very long time…
To backup one single file of 60GB via ssh it use no more than 2hours… (with same hardware and network).
I’ve increased CACHESIZE but without results…
P.S.: to transfer files between servers I use nfs to mount lessfs partition and rdiff-backup. (same results if I use rsync, scp or similar)
“Centos the kernel does not support >4096 page size”
What’s the consequences of that?