FSCK Part 2/2 - Non-contiguous inodes query
Having got to the point where I could run the fsck binary, I did so using the command: fsck -v -p -c -f /dev/hdb1 etc as listed in my /etc/fstab file, besides a few like the swap, usbdev. Having the verbose mode on allowed me to see what details were displayed for each partition I ran it on. I noticed that on all the drives fsck'd they all had non=contiguous inodes percentages of anything from 0 to 15% if memory serves me. Considering all the partitions I did were ext3 how come the nc inodes are so high? I thought with ext3 being a journaling fs they would all be 0? am I right in assuming that the nc inodes is similar to how much the partition is fragmented? I looked on Google for 'What are non-contiguous inodes' but there were just too many results not giving a direct answer to the question. How do I recover the nc inodes or get the %'s to 0? -- The Little Helper ======================================================================== Hylton Conacher - Licenced ex-Windows user (apart from Quicken) Registered Linux user # 229959 at http://counter.li.org Currently using SuSE 9.0 Professional with KDE 3.1 ========================================================================
-----Original Message-----
From: "Hylton Conacher (ZR1HPC)"
Having got to the point where I could run the fsck binary, I did so using the command: fsck -v -p -c -f /dev/hdb1 etc as listed in my /etc/fstab file, besides a few like the swap, usbdev.
Having the verbose mode on allowed me to see what details were displayed for each partition I ran it on. I noticed that on all the drives fsck'd they all had non=contiguous inodes percentages of anything from 0 to 15% if memory serves me.
Considering all the partitions I did were ext3 how come the nc inodes are so high? I thought with ext3 being a journaling fs they would all be 0? am I right in assuming that the nc inodes is similar to how much the partition is fragmented? I looked on Google for 'What are non-contiguous inodes' but there were just too many results not giving a direct answer to the question.
How do I recover the nc inodes or get the %'s to 0?
You don't ever get them to 0%. What makes 15% so high? With todays high speed disk drives fragmentation is not such a problem unless using MS that does not continously defrag the drive. That's why you don't readily see defrag software for unix systems. If you have fixed the partition problems using fsck just enjoy your system. Ken
On Saturday 06 Mar 2004 11:30 am, Hylton Conacher (ZR1HPC) wrote: <SNIP>
I noticed that on all the drives fsck'd they all had non=contiguous inodes percentages of anything from 0 to 15% if memory serves me.
Considering all the partitions I did were ext3 how come the nc inodes are so high?
15% isn't so high, especially when you consider a heavily used FAT fs can get well over 50% quite quickly...
I thought with ext3 being a journaling fs they would all be 0?
The journalling has nothing to do with nc-nodes SFAIK.
am I right in assuming that the nc inodes is similar to how much the partition is fragmented?
Yes. <SNIP>
How do I recover the nc inodes or get the %'s to 0?
There's no need at all. The *nix fs's are inherently more efficient than M$ fs's so the level of fragmentation has much less effect on the performance. Also, defragmentation occurs in the background - have you noticed the drive(s) 'chuntering' at times of low activity? Overall, you don't need to worry about nc inodes unless you see them running out of control (which would be unlikely, I think.) Dylan -- "They that can give up essential liberty to obtain a little temporary safety, deserve neither liberty nor safety." -Benjamin Franklin
Dylan wrote:
On Saturday 06 Mar 2004 11:30 am, Hylton Conacher (ZR1HPC) wrote: <SNIP>
I noticed that on all the drives fsck'd they all had non=contiguous inodes percentages of anything from 0 to 15% if memory serves me.
Considering all the partitions I did were ext3 how come the nc inodes are so high?
15% isn't so high, especially when you consider a heavily used FAT fs can get well over 50% quite quickly... Tnx Dylan, Ken and any others I missed.
-- The Little Helper ======================================================================== Hylton Conacher - Licenced ex-Windows user (apart from Quicken) Registered Linux user # 229959 at http://counter.li.org Currently using SuSE 9.0 Professional with KDE 3.1 ========================================================================
participants (3)
-
Dylan
-
Hylton Conacher (ZR1HPC)
-
Ken Schneider