Project

General

Profile

Bug #21712

ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer lengths mismatch expected 0x1a2c1 != actual 0x0

Added by Sage Weil over 6 years ago. Updated about 3 years ago.

Status:
Resolved
Priority:
High
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
luminous
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

2017-10-05T21:05:47.120 INFO:teuthology.orchestra.run.smithi004.stdout:[ RUN      ] ObjectStore/StoreTest.Synthetic/1
2017-10-05T21:05:48.259 INFO:teuthology.orchestra.run.smithi004.stderr:seeding object 0
2017-10-05T21:05:49.808 INFO:teuthology.orchestra.run.smithi004.stderr:seeding object 500
2017-10-05T21:05:50.352 INFO:teuthology.orchestra.run.smithi004.stderr:Op 0
2017-10-05T21:05:50.352 INFO:teuthology.orchestra.run.smithi004.stderr:available_objects: 984 in_flight_objects: 16 total objects: 1000 in_flight 16
2017-10-05T21:06:03.960 INFO:teuthology.orchestra.run.smithi004.stdout:--- buffer lengths mismatch expected 0x1a2c1 != actual 0x0

/a/yuriw-2017-10-05_19:50:51-rados-wip-yuri-testing3-2017-10-04-1630-distro-basic-smithi/1706936

Related issues

Related to RADOS - Bug #21818: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer content mismatch Resolved 10/06/2017
Copied to Ceph - Backport #21731: luminous: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer lengths mismatch expected 0x1a2c1 != actual 0x0 Rejected

History

#1 Updated by Sage Weil over 6 years ago

it appears to be a problem with 'zero' semantics:

   -19> 2017-10-05 21:06:03.896257 7ff0d773c700 15 filestore(filestore.test_temp_dir) _zero(3546): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 107201~0
   -18> 2017-10-05 21:06:03.896271 7ff0d773c700 20 filestore(filestore.test_temp_dir) _zero(3602): falling back to writing zeros
   -17> 2017-10-05 21:06:03.896276 7ff0d773c700 15 filestore(filestore.test_temp_dir) _write(3499): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 107201~0
   -16> 2017-10-05 21:06:03.896304 7ff0d5f39700 15 filestore(filestore.test_temp_dir) read(3225): 555.0_head/#555:b3000000:::OBJ_460:3aa1e1c3# 0~190734
   -15> 2017-10-05 21:06:03.896345 7ff0d773c700 10 filestore(filestore.test_temp_dir) _write(3540): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 107201~0 = 0
   -14> 2017-10-05 21:06:03.896356 7ff0d773c700 20 filestore(filestore.test_temp_dir) _zero(3616): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 107201~0 = 0
...
    -3> 2017-10-05 21:06:03.896662 7ff0d5f39700 15 filestore(filestore.test_temp_dir) read(3225): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 0~107201
    -2> 2017-10-05 21:06:03.896695 7ff0d5f39700 10 filestore(filestore.test_temp_dir) read(3280): 555.0_head/#555:d5000000:::OBJ_938:4756f8d2# 0~0/107201

#2 Updated by Sage Weil over 6 years ago

The problem is that a 0-length write does not extend the size, but a 0-length zero is changing the object size.

#4 Updated by Sage Weil over 6 years ago

  • Backport set to luminous

#5 Updated by Sage Weil over 6 years ago

  • Status changed from Fix Under Review to Pending Backport

#6 Updated by Nathan Cutler over 6 years ago

  • Copied to Backport #21731: luminous: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer lengths mismatch expected 0x1a2c1 != actual 0x0 added

#7 Updated by Kefu Chai over 6 years ago

  • Copied to Bug #21818: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer content mismatch added

#8 Updated by Nathan Cutler over 6 years ago

@Kefu: Changing the #21818 link to "Related" instead of "Copied" since I'm guessing #21818 is a separate bug; not a backport?

#9 Updated by Nathan Cutler over 6 years ago

  • Copied to deleted (Bug #21818: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer content mismatch)

#10 Updated by Nathan Cutler over 6 years ago

  • Related to Bug #21818: ceph_test_objectstore fails ObjectStore/StoreTest.Synthetic/1 (filestore) buffer content mismatch added

#11 Updated by Kefu Chai over 6 years ago

@Nathan, thanks for fixing this. i didn't realize that "Copied" implied "backport", i just used "copy" as a handy tool to create a new bug.

#12 Updated by Kefu Chai over 6 years ago

  • Status changed from Pending Backport to Fix Under Review

#13 Updated by Nathan Cutler over 6 years ago

i didn't realize that "Copied" implied "backport"

Copy is used to make the backport tracker issues, and sometimes it's done manually and the person doing it forgets to change the tracker (of the copy) from Bug to Backport. That was my assumption when I saw you had copied, but then I looked closer and realized you were creating a new bug. So yes, the copy function does imply a backport, but probably just for me, so don't let that stop you from using it :-)

#14 Updated by Sage Weil over 6 years ago

  • Status changed from Fix Under Review to Pending Backport

#15 Updated by Nathan Cutler about 3 years ago

  • Status changed from Pending Backport to Resolved

While running with --resolve-parent, the script "backport-create-issue" noticed that all backports of this issue are in status "Resolved" or "Rejected".

Also available in: Atom PDF