Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1091701

Summary: [SCALE] Disk creation on local disk takes too long
Product: Red Hat Enterprise Virtualization Manager Reporter: Kevin Alon Goldblatt <kgoldbla>
Component: vdsmAssignee: Tal Nisan <tnisan>
Status: CLOSED WORKSFORME QA Contact: Aharon Canan <acanan>
Severity: high Docs Contact:
Priority: unspecified    
Version: 3.4.0CC: amureini, bazulay, gklein, iheim, kgoldbla, lpeer, scohen, tnisan, yeylon
Target Milestone: ---Flags: amureini: Triaged+
Target Release: 3.5.0   
Hardware: x86_64   
OS: Linux   
Whiteboard: storage
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-07-20 05:08:27 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Storage RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1091712    
Attachments:
Description Flags
vdsm and engine logs none

Description Kevin Alon Goldblatt 2014-04-27 12:14:08 UTC
Created attachment 890214 [details]
vdsm and engine logs

Description of problem:
Creating a 125G local disk took aroung 2 hours to complete and fail

Version-Release number of selected component (if applicable):
rhevm-3.4.0-0.12.beta2.el6ev.noarch
vdsm-4.14.6-0.1.beta3.el6ev.x86_64

How reproducible:
1/1 takes very long

Steps to Reproduce:
1. Create vm with 3 disks (2 thin, 1 preallocated)
2. Free space displays 135G but actually should be less as the free space is updated slowly.
3. Selected to create a preallocated disk of 125G (actual space should be less)
3. After about 1 hour and 40 min the operation failed 

Actual results:
The create disk operation failed after 1 hour and 40 min

Expected results:
The operation should not have taken this long to completed even if there was a disk space issue

Additional info:
isk takes too long to create: 27/4/2014
---------------------------------------------------------------
engine===========
2014-04-27 10:49:33,657 INFO  [org.ovirt.engine.core.bll.AddDiskCommand] (ajp-/127.0.0.1:8702-6) [5047e196] Lock Acquired to object EngineLock [exclusiveLocks= , sharedLocks= key: 4be5ee20-c3a1-45b0-932d-ea6203cc8762 value: VM
2014-04-27 10:49:33,687 INFO  [org.ovirt.engine.core.bll.AddDiskCommand] (ajp-/127.0.0.1:8702-6) [5047e196] Running command: AddDiskCommand internal: false. Entities affected :  ID: 4be5ee20-c3a1-45b0-932d-ea6203cc8762 Type: VM,  ID: 11031344-4f1a-452f-bca8-8e9c6b68596b Type: Storage
2014-04-27 10:49:34,529 INFO  [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (ajp-/127.0.0.1:8702-6) [4d572807] Correlation ID: 5047e196, Job ID: 50a581b4-b25b-47bd-bef9-98f7b359fb49, Call Stack: null, Custom Event ID: -1, Message: Add-Disk operation of test4_Disk6 was initiated on VM test4 by admin.
2014-04-27 12:25:35,816 ERROR [org.ovirt.engine.core.bll.AddDiskCommand] (org.ovirt.thread.pool-4-thread-20) [5047e196] Ending command with failure: org.ovirt.engine.core.bll.AddDiskCommand
2014-04-27 12:25:35,923 INFO  [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (org.ovirt.thread.pool-4-thread-20) Correlation ID: 5047e196, Job ID: 50a581b4-b25b-47bd-bef9-98f7b359fb49, Call Stack: null, Custom Event ID: -1, Message: Operation Add-Disk failed to complete.

Comment 1 Allon Mureinik 2014-04-27 14:12:19 UTC
On the same storage, how long does it take to dd 10GB from /dev/zero to a file?

Comment 2 Kevin Alon Goldblatt 2014-04-28 11:39:33 UTC
(In reply to Allon Mureinik from comment #1)
> On the same storage, how long does it take to dd 10GB from /dev/zero to a
> file?

[root@nott-vds2 mnt]# time dd if=/dev/zero of=/mnt/10g bs=10M count=10000
718+0 records in
718+0 records out
7528775680 bytes (7.5 GB) copied, 238.915 s, 31.5 MB/s
1340+0 records in
1340+0 records out
14050918400 bytes (14 GB) copied, 498.018 s, 28.2 MB/s

Comment 3 Allon Mureinik 2014-04-29 07:26:52 UTC
perallocation 125G in 1:40 hours is roughly 21 MB/s. It's slightly slower than how your storage performs, but not dramatically.

May be worth researching as part of a scale initiative, but IMHO, not urgent for 3.4.

Comment 4 Sean Cohen 2014-06-11 08:30:40 UTC
Kevin,
Please reproduce on a non local storage and provide results, I am not sure it is a scale issue, rather than a slow local medium.
Sean

Comment 5 Kevin Alon Goldblatt 2014-06-15 13:13:02 UTC
I ran the scenario on a no local storage and the system behave correclty as follows:

1) All attempts to create a disk larger than the available storage failed almost immediately

2) Creating a 125G disk took less than a minute

Comment 6 Allon Mureinik 2014-06-15 15:41:00 UTC
(In reply to Kevin Alon Goldblatt from comment #5)
> I ran the scenario on a no local storage and the system behave correclty as
> follows:
> 
> 1) All attempts to create a disk larger than the available storage failed
> almost immediately
> 
> 2) Creating a 125G disk took less than a minute
With pre-allocation?

Comment 7 Kevin Alon Goldblatt 2014-06-17 07:22:25 UTC
Yes,

I created both Virt-IO and Virt-IO SCSI Preallocated disks of 125G and they were created in less than a minute

Comment 8 Tal Nisan 2014-07-08 15:12:05 UTC
I've created a 125GB disk on a local domain, took ~18 minutes, create a 125GB with dd from /dev/zero on the same host, took ~18 minutes.
This seems to work correctly for me

Comment 9 Sean Cohen 2014-07-20 05:08:27 UTC
Closing per comment #8
Feel free to re-open if you have new indication
Sean