Re: [Ocfs2-users] kworker high disk io

[Thomas Manninger]

After waiting about 20 minutes, i get the following output:

echo "locate <W00000000000000015909850d471ca3>"  | debugfs.ocfs2 /dev/sdb
debugfs.ocfs2 1.8.6
debugfs: locate <W00000000000000015909850d471ca3>
    22612357    //orphan_dir:0002/0000000001590985/

 

Why are deleted files are locked?

 

Gesendet: Mittwoch, 19. April 2017 um 09:43 Uhr
Von: "Thomas Manninger" <dbgtmas...@gmx.at>
An: ocfs2-users@oss.oracle.com
Betreff: [Ocfs2-users] kworker high disk io

Hi,

 

i have a problem with my ocfs2 cluster filesystem. I have 3 nodes, running with CentOS 7.

 

The cluster runs since 1 year without problems. Last week, one colleague deletes about 100.000 files on the Filesystem with "rm *" in some folders.

Since that, the kworker io is about -50-80% usage (iotop), also iostat have the same value.

 

If i list all "fs_locks", i get about 190.000 entries:

echo "fs_locks"  | debugfs.ocfs2 -n /dev/sdb | grep "Lockres" | wc -l

 

That are too many locks?

 

As example, i get the following output:

Lockres: M00000000000000062e723300000000  Mode: Exclusive
Flags: Initialized Attached
RO Holders: 0  EX Holders: 0
Pending Action: None  Pending Unlock Action: None
Requested Mode: Exclusive  Blocking Mode: Invalid
PR > Gets: 16  Fails: 0    Waits Total: 657us  Max: 654us  Avg: 41072ns
EX > Gets: 2  Fails: 0    Waits Total: 322us  Max: 322us  Avg: 161398ns
Disk Refreshes: 0

 

If i try to get the filename, the command hangs and i never get some output:

echo "locate <M00000000000000062e723300000000>"  | debugfs.ocfs2 -n /dev/sdb

 

How can i solve this problem?

 

Best regards,

Thomas Manninger

_______________________________________________ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com https://oss.oracle.com/mailman/listinfo/ocfs2-users

_______________________________________________
Ocfs2-users mailing list
Ocfs2-users@oss.oracle.com
https://oss.oracle.com/mailman/listinfo/ocfs2-users