Project

General

Profile

Actions
Copy link

Bug #9046

closed

Limiting the pool object quota stops the IO, however IO does not restart if we reset the pool object quota to a higher number

Added by Hirak Mazumder over 9 years ago. Updated over 9 years ago.

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

0%

Source:
Community (user)
Tags:
Backport:
Regression:
Severity:
2 - major
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Issue Title: Limiting the pool object quota stops the IO, however IO does not restart if we rest the pool object quota to a higher number

Steps to recreate:

1. Create a single node cluster
2. Create 8 OSDs on the disks connected to Igbox
3. Create a single monitor
4. Create a pool with 1024 PGs
5. Create 10 RBDs on the pool
6. Map the RBDs to the client
7. Run IO using FIO for an hour and note the object count
8. Reset the pool quota to hold 1000 objects
9. Note the IO status
10. Reset the pool quota to hold 300K objects
11. Note the IO status

Observations:
Noticed that in Step 7 object count is 250K
After resetting the pool quota to 1000, noticed that the client IO stopped which is expected
Noticed that after resetting the pool quota to 300K objects does not restart the IO.

Expected Results: IOs should have been restarted.

Files

Download all files
logs.zip (4.42 MB) logs.zip Hirak Mazumder, 08/08/2014 10:31 AM
Logs.txt (29.6 KB) Logs.txt Hirak Mazumder, 08/08/2014 10:31 AM
Actions
Copy link
 #1

Updated by Hirak Mazumder over 9 years ago

Hirak Mazumder wrote:

Issue Title: Limiting the pool object quota stops the IO, however IO does not restart if we rest the pool object quota to a higher number

Steps to recreate:

1. Create a single node cluster
2. Create 8 OSDs on the disks connected to Igbox
3. Create a single monitor
4. Create a pool with 1024 PGs
5. Create 10 RBDs on the pool
6. Map the RBDs to the client
7. Run IO using FIO for an hour and note the object count
8. Reset the pool quota to hold 1000 objects
9. Note the IO status
10. Reset the pool quota to hold 300K objects
11. Note the IO status

Observations:
Noticed that in Step 7 object count is 250K
After resetting the pool quota to 1000, noticed that the client IO stopped which is expected
Noticed that after resetting the pool quota to 300K objects does not restart the IO.

Expected Results: IOs should have been restarted.

Actions
Copy link
 #2

Updated by Hirak Mazumder over 9 years ago

Please let me know if there is any update on this issue?

Actions
Copy link
 #3

Updated by Sage Weil over 9 years ago

  • Priority changed from High to Urgent
Actions
Copy link
 #4

Updated by Sage Weil over 9 years ago

  • Status changed from New to Resolved

This bug was fixed in the 3.14 kernel. There have been several other fixes since then as well... your best bet is to use the v3.16 kernel. See https://wiki.ubuntu.com/Kernel/MainlineBuilds

Actions
Copy link

Also available in: Atom PDF