[Bugs] [Bug 1242504] New: [Data Tiering]: Frequency Counters of un-selected file in the DB wont get clear after a promotion/demotion cycle

bugzilla at redhat.com bugzilla at redhat.com
Mon Jul 13 12:56:03 UTC 2015


https://bugzilla.redhat.com/show_bug.cgi?id=1242504

            Bug ID: 1242504
           Summary: [Data Tiering]: Frequency Counters of un-selected file
                    in the DB wont get clear after a promotion/demotion
                    cycle
           Product: GlusterFS
           Version: mainline
         Component: tiering
          Severity: high
          Assignee: bugs at gluster.org
          Reporter: josferna at redhat.com
        QA Contact: bugs at gluster.org
                CC: bugs at gluster.org



Description of problem:

Change Time Recorder increments the write/read frequency counters on a read or
write of a file, if the "features.record-counters" is "on". It is the
responsibility of the tiering migrator to reset these counters to zero for
un-selected files to reset them to zero as frequency counters are function of
promotion/Demotion cycles. If the counters are not set to zero then,

1) the counters may overflow in the DB
2) The file may be wrongly promoted or demoted.

To test if the counters are cleared for un-selected after a promotion/demotion
cycle is by executing the following sqlite3 query.

$>  echo "select GF_ID, WRITE_FREQ_CNTR,READ_FREQ_CNTR from GF_FILE_TB;" |
sqlite3 <brick_path>/.glusterfs/<brick_name>.db;


Version-Release number of selected component (if applicable):


How reproducible:

1) Create a dist-rep volume 
2) Start volume
3) attach a dist-rep hot tier
4) gluster volume set features.ctr-enabled on ; gluster volume 
features.record-counters on;
5) gluster volume set test cluster.read-freq-threshold <same_value> 
6) gluster volume set test cluster.write-freq-threshold <same_value>
7) create some files, heat up some files and leave other files so that cold
files get demoted.
8) After a demotion cycle execute the above mentioned sqlite query on hot brick
dbs, you will observe that the counters are not cleared.


Actual results:

The Frequecy counters of un-selected files are NOT set to zero after a
promotion/demotion cycle.


Expected results:

The Frequecy counters of un-selected files should be set to zero after a
promotion/demotion cycle.

Additional info:

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


More information about the Bugs mailing list