[Bugs] [Bug 1785520] New: performance/io-cache: After “performance/io-cache: update pages with write data” patch commit and disable md-cache sometimes that gcc compilation file can not execute right now.

bugzilla at redhat.com bugzilla at redhat.com
Fri Dec 20 06:12:15 UTC 2019


https://bugzilla.redhat.com/show_bug.cgi?id=1785520

            Bug ID: 1785520
           Summary: performance/io-cache: After “performance/io-cache:
                    update pages with write data” patch commit and disable
                    md-cache sometimes that gcc  compilation file can not
                    execute right now.
           Product: GlusterFS
           Version: 6
          Hardware: x86_64
                OS: Linux
            Status: NEW
         Component: io-cache
          Severity: high
          Assignee: bugs at gluster.org
          Reporter: 13965432176 at 163.com
                CC: bugs at gluster.org
  Target Milestone: ---
           Link ID: Red Hat Bugzilla 1659869,Gluster.org Gerrit 1659869
    Classification: Community



Description of problem:
The patch SHA-1: d50f22e6ae410fdcde573b6015b97dc1573bbb7e discripted that
"performance/io-cache: update pages with write data” . How after that patch
commit and disable md-cache sometimes that gcc  compilation file can not
execute right now.

Version-Release number of selected component (if applicable):
gluster version 6, io-cache

Steps to Reproduce:
1.gluster v create testvol_distributed
host1:/bricks/brick1/testvol_distributed_brick0
host2:/bricks/brick1/testvol_distributed_brick0
host3:/bricks/brick1/testvol_distributed_brick0
2.gluster v set testvol_distributed performance.md-cache-pass-through true
3.gluster v start testvol_distributed 
4.mount -t glusterfs 127.0.0.1:/testvol_distributed /mnt/testvol_dispersed
5.gcc stat.c -o stat1 -lm ; ./stat1 smcomp.time
-bash: ./stat1: cannot execute binary file
6.we should execute stat1 right now. we maybe not reproduce this problem so 

Actual results:
can execute binary file

Expected results:
can execute binary file

Additional info:
The reason is not do ioc_inode_flush or update io cache page failed after
ioc_writev_cbk. The patch commit ioc_inode_flush used by steps
mdc_inode_iatt_set_validate->inode_invalidate->ioc_invalidate->ioc_inode_flush
.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.


More information about the Bugs mailing list