File system performance

circus78 · 12-01-2014, 03:22 AM

Hi,
I have big performance issue on my file system.
It's a Coraid attached storage, size is 1.4 TB (occupation 98%).
file system is ext4, mounted with noatime.

I have no network issues (server and Coraid are directly attached and I have no error on ethernet interface).

My question is: could fsck improve performance?

Code:

# tune2fs -l /dev/etherd/e1.1p1
tune2fs 1.41.12 (17-May-2010)
Filesystem volume name:   <none>
Last mounted on:          /var/spool/pop
Filesystem UUID:          9177ca3b-d41e-4b26-965e-20aead66dd68
Filesystem magic number:  0xEF53
Filesystem revision #:    1 (dynamic)
Filesystem features:      has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize
Filesystem flags:         signed_directory_hash
Default mount options:    (none)
Filesystem state:         clean
Errors behavior:          Continue
Filesystem OS type:       Linux
Inode count:              91578368
Block count:              366286008
Reserved block count:     18314300
Free blocks:              73660901
Free inodes:              78532783
First block:              0
Block size:               4096
Fragment size:            4096
Reserved GDT blocks:      936
Blocks per group:         32768
Fragments per group:      32768
Inodes per group:         8192
Inode blocks per group:   512
Flex block group size:    16
Filesystem created:       Wed Nov 14 10:54:50 2012
Last mount time:          Mon May 26 11:49:12 2014
Last write time:          Mon May 26 11:49:12 2014
Mount count:              9
Maximum mount count:      30
Last checked:             Tue Dec 18 18:03:41 2012
Check interval:           15552000 (6 months)
Next check after:         Sun Jun 16 19:03:41 2013
Lifetime writes:          4021 GB
Reserved blocks uid:      0 (user root)
Reserved blocks gid:      0 (group root)
First inode:              11
Inode size:               256
Required extra isize:     28
Desired extra isize:      28
Journal inode:            8
First orphan inode:       29508801
Default directory hash:   half_md4
Directory Hash Seed:      36e88180-9f85-4c62-a705-04c4dbeef8bc
Journal backup:           inode blocks

How can I predict how long it will take?
Thank you very much

linosaurusroot · 12-01-2014, 06:12 AM

You haven't said anything about how you observed (let alone measured) the performance. Knowing where the time goes and what you are doing during poor performance should inform your planning.

One thing you can do is limit the number of files per directory. Also for any directory that has had an excessive number of files in it replace it (by making a new empty directory; move the contents to the new one; finally rmdir the old one).

circus78 · 12-01-2014, 08:38 AM

Quote:

Originally Posted by linosaurusroot

You haven't said anything about how you observed (let alone measured) the performance. Knowing where the time goes and what you are doing during poor performance should inform your planning.

One thing you can do is limit the number of files per directory. Also for any directory that has had an excessive number of files in it replace it (by making a new empty directory; move the contents to the new one; finally rmdir the old one).

Hi linosaurusroot,
I used "top" command and I see very high I/O wait.
This partition contains mail spool of many users (> 9000).
Mail daemon is dovecot, webmail is Roundcube.
Thank you again.

pan64 · 12-01-2014, 09:30 AM

probably adding some more space will help you. (or clean up this filesystem). I do not think fsck will help you.