Lessfs now supports transactions which eliminates the need for fsck after a crash. Also new is the automatic execution of a custom script when the file system has less then a certain amount of free space left.
-
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
Great! Time to upgrade I guess.
Just have to wait until lessfsck is done. It it running for more than 24h now. I defenitly need to upgrade to newer hardware
Keep up the good work!
Bas
Yeah! what a drop
from 1.0.1 to 1.0.4 ! This is a greaaaaaaaaaaat news ! Lessfsck was really a pain taking endless time to get production servers back online ! this is an amazing news :p
Gonna upgrade the version on test, then prod server asap :p
Thank you mark! This is THE good news of the week
Best
Any news about getting compression and dedupe ratio from lessfs?
I’ve made a “shell script” to get the compression ratio and amount of deduplication from lessfs:
grep ‘lessfs\[8118\]’ /var/log/messages.log | grep -Eo ‘size .* bytes \: lzo compressed bytes .* and .* duplicate blocks’ | awk ‘ { SUM += $2 ; SUMCP += $8 ; DUP += $10 } END { print “total size ” SUM/1024/1024/1024 “GiB, compressed size ” SUMCP/1024/1024/1024 “GiB (ratio 1 : ” SUM/SUMCP “) with ” DUP ” duplicate blocks (” DUP*64/1024/1024 “GiB deduplicated)” }’
change lessfs\[8118\] to the one you have in use and the DUP*64 is for 64KiB block size, change it if you use different
Hi mark, just a question, have tried to unplug a drive using a mounted volume off lessfs, and great, while remounting it gets back really quick.
But, what makes change on the rollback date ? here is still “Thu Jan 1 01:00:00 1970″, even after writing a file, unpluging drive, mouting it back again.
Mar 5 22:49:22 localhost lessfs[28892]: The selected data store is tokyocabinet.
Mar 5 22:49:22 localhost lessfs[28892]: Lessfs transaction support is disabled.
Mar 5 22:49:22 localhost lessfs[28892]: Lessfs uses a 24 bytes long hash.
Mar 5 22:49:22 localhost lessfs[28892]: Lessfs fsync does not sync the databases to the disk when fsync is called on an inode
Mar 5 22:49:22 localhost lessfs[28892]: cache 12288 data blocks
Mar 5 22:49:22 localhost lessfs[28892]: The tiger hash has been selected.
Mar 5 22:49:23 localhost lessfs[28896]: Lessfs has not been unmounted cleanly.
Mar 5 22:49:23 localhost lessfs[28896]: Rollback to : Thu Jan 1 01:00:00 1970
How can i set the new rollback point ? something like “sync” ?
Well, the sync command made the changes are kept in the volume (don’t know if there’s a link) but the “rollback to” date information in the log is allways “Thu Jan 1 01:00:00 1970″. So i think it’s just something in the way logs are handled.
Great news!
I’ve been thinking about using lessfs together with Samba shadow copy support to give users a native and easy to use access to their backups, problem is, we need a fairly complex security model that needs ACLs just so users couldn’t access files they don’t normally have access to.
Would it be possible for you to add support for POSIX ACLs?
(the extended attributes would be nice too, but they are not critical)
Thank you for your work
Hubert
The rollback logging is a bug.
You did not enable transactions.
See the second line of the logging
Hi Mark,
I just noticed the same logging thing as Alex talked about.
When I checked my syslog, I also noticed the “Last Used” and “Filesystem is clean” entries come with a critical severity. Seems more like an info or notice severity to me.
All in all the rest works great.
I have just one small question: Now that I have the volume running, I see a few databases, apart from the big one. You know like hardlinks and stuff.
These small files are needed I suppose, but what happens if they are gone? are they easy to rebuild?
Cheers, Bas
I’ll change the logging thing as soon as I have time.
About ‘damaged / missing databases’.
Some databases can be reconstructed, like hardlinks, symlinks, blockusage.
Others metdata, fileblock, blockdata can not be reconstructed when you delete them.
That will be great.
And thank you for the elaboration on the extra files. I’ll keep them safe
Greetings, Bas
Enabling transactions will provide protection against most cases of database corruption. I strongly recommend this new feature.
Groet,
Mark
Hi Mark,
New feature enabled now of course !! this is a must have
shame that i hadn’t seen the option before reporting information. But, in a way, this is how we learn how to use it.