Bug 49717

Summary:	no space left on device error
Product:	[Retired] Red Hat Linux	Reporter:	Don Smith <sdonald>
Component:	kernel	Assignee:	Arjan van de Ven <arjanv>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Aaron Brown <abrown>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	7.3
Target Milestone:	---
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2001-08-09 21:53:26 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Don Smith 2001-07-23 14:43:38 UTC

From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)

Description of problem:
Bonnie (www.textuality.com/bonnie) fails on a write() call while writing 
on a RAID 1 device. The message printed by the perror() call made by 
bonnie after the failed write() is "No space left on device." However, df 
reports that there is plenty of space left on the device (shown below).



How reproducible:
Sometimes

Steps to Reproduce:
1. Run Bonnie to write a large file (at least 512MB) on a RAID 1 device.
2. Bonnie should fail as described.
	

Additional info:

IBM ServeRAID 4L controller on a dual P3-700 w/1024MB RAM.

Bonnie output:
 File '/temp/Bonnie.1185', size: 536870912
 Writing with putc()...done
 Rewriting...done
 Writing intelligently...Bonnie: drastic I/O error (write(2)): No space 
left on device

df reports the following:
/dev/sdb1
 1k-blocks: 1031608
 Used: 506804
 Available: 472400
 Use%: 52%
 Mounted on: /temp

The actual size of the file when Bonnie dies:
/temp/Bonnie.1185
 size: 501649408

We reproduced this once - failed much earlier into the file (324444160 
bytes).

Comment 1 Arjan van de Ven 2001-07-23 16:42:40 UTC

Smells like the VM mistuning gives -ENOMEM and bonnie translates that as "disk
full"....

Comment 2 Glen Foster 2001-07-23 21:17:22 UTC

This defect is considered SHOULD-FIX for Fairfax.

Comment 3 Arjan van de Ven 2001-08-03 14:04:40 UTC

Beta3 should have this fixed at the cost of performance-loss. We're trying
to undo the performance-loss without loosing the fix. Could you verify this ?

Comment 4 Don Smith 2001-08-09 21:53:21 UTC

This failure no longer occurs in beta 3.