diff mbox

[2/3] btrfs: add replace missing and replace RAID 5/6 to profile configs

Message ID 6d332e954da29e8418689cb49294a68da2fa5376.1437683940.git.osandov@fb.com (mailing list archive)
State New, archived
Headers show

Commit Message

Omar Sandoval July 23, 2015, 8:51 p.m. UTC
Replacing and scrubbing RAID 5/6 is now supported on Btrfs. Enable it in
_btrfs_get_profile_configs while making it more generic to also support
replace missing.

Signed-off-by: Omar Sandoval <osandov@fb.com>
---
 common/rc | 96 ++++++++++++++++++++++++++++++++-------------------------------
 1 file changed, 49 insertions(+), 47 deletions(-)

Comments

Eryu Guan July 24, 2015, 3:03 a.m. UTC | #1
On Thu, Jul 23, 2015 at 01:51:50PM -0700, Omar Sandoval wrote:
> Replacing and scrubbing RAID 5/6 is now supported on Btrfs. Enable it in
> _btrfs_get_profile_configs while making it more generic to also support
> replace missing.
> 
> Signed-off-by: Omar Sandoval <osandov@fb.com>

Looks great! Tested with default configs and user-defined configs.

Reviewed-by: Eryu Guan <eguan@redhat.com>

> ---
>  common/rc | 96 ++++++++++++++++++++++++++++++++-------------------------------
>  1 file changed, 49 insertions(+), 47 deletions(-)
> 
> diff --git a/common/rc b/common/rc
> index 610045eab304..3e6fdb6ebcfa 100644
> --- a/common/rc
> +++ b/common/rc
> @@ -2748,60 +2748,62 @@ _btrfs_get_profile_configs()
>  		return
>  	fi
>  
> -	# no user specified btrfs profile configs, export the default configs
>  	if [ -z "$BTRFS_PROFILE_CONFIGS" ]; then
> -		# default configs
> -		_btrfs_profile_configs=(
> -			"-m single -d single"
> -			"-m dup -d single"
> -			"-m raid0 -d raid0"
> -			"-m raid1 -d raid0"
> -			"-m raid1 -d raid1"
> -			"-m raid10 -d raid10"
> -			"-m raid5 -d raid5"
> -			"-m raid6 -d raid6"
> +		# Default configurations to test.
> +		local configs=(
> +			"single:single"
> +			"dup:single"
> +			"raid0:raid0"
> +			"raid1:raid0"
> +			"raid1:raid1"
> +			"raid10:raid10"
> +			"raid5:raid5"
> +			"raid6:raid6"
>  		)
> +	else
> +		# User-provided configurations.
> +		local configs=(${BTRFS_PROFILE_CONFIGS[@]})
> +	fi
>  
> -		# remove dup/raid5/raid6 profiles if we're doing device replace
> -		# dup profile indicates only one device being used (SCRATCH_DEV),
> -		# but we don't want to replace SCRATCH_DEV, which will be used in
> -		# _scratch_mount/_check_scratch_fs etc.
> -		# and raid5/raid6 doesn't support replace yet
> +	_btrfs_profile_configs=()
> +	for cfg in "${configs[@]}"; do
> +		local supported=true
> +		local profiles=(${cfg/:/ })
>  		if [ "$1" == "replace" ]; then
> -			_btrfs_profile_configs=(
> -				"-m single -d single"
> -				"-m raid0 -d raid0"
> -				"-m raid1 -d raid0"
> -				"-m raid1 -d raid1"
> -				"-m raid10 -d raid10"
> -				# add these back when raid5/6 is working with replace
> -				#"-m raid5 -d raid5"
> -				#"-m raid6 -d raid6"
> +			# We can't do replace with these profiles because they
> +			# imply only one device ($SCRATCH_DEV), and we need to
> +			# keep $SCRATCH_DEV around for _scratch_mount
> +			# and _check_scratch_fs.
> +			local unsupported=(
> +				"single"
> +				"dup"
>  			)
> +		elif [ "$1" == "replace-missing" ]; then
> +			# We can't replace missing devices with these profiles
> +			# because there isn't enough redundancy.
> +			local unsupported=(
> +				"single"
> +				"dup"
> +				"raid0"
> +			)
> +		else
> +			local unsupported=()
>  		fi
> -		export _btrfs_profile_configs
> -		return
> -	fi
> -
> -	# parse user specified btrfs profile configs
> -	local i=0
> -	local cfg=""
> -	for cfg in $BTRFS_PROFILE_CONFIGS; do
> -		# turn "metadata:data" format to "-m metadata -d data"
> -		# and assign it to _btrfs_profile_configs array
> -		cfg=`echo "$cfg" | sed -e 's/^/-m /' -e 's/:/ -d /'`
> -		_btrfs_profile_configs[$i]="$cfg"
> -		let i=i+1
> -	done
> -
> -	if [ "$1" == "replace" ]; then
> -		if echo ${_btrfs_profile_configs[*]} | grep -q raid[56]; then
> -			_notrun "RAID5/6 doesn't support btrfs device replace yet"
> -		fi
> -		if echo ${_btrfs_profile_configs[*]} | grep -q dup; then
> -			_notrun "Do not set dup profile in btrfs device replace test"
> +		for unsupp in "${unsupported[@]}"; do
> +			if [ "${profiles[0]}" == "$unsupp" -o "${profiles[1]}" == "$unsupp" ]; then
> +			     if [ -z "$BTRFS_PROFILE_CONFIGS" ]; then
> +				     # For the default config, just omit it.
> +				     supported=false
> +			     else
> +				     # For user-provided config, don't run the test.
> +				     _notrun "Profile $unsupp not supported for $1"
> +			     fi
> +			fi
> +		done
> +		if "$supported"; then
> +			_btrfs_profile_configs+=("-m ${profiles[0]} -d ${profiles[1]}")
>  		fi
> -	fi
> +	done
>  	export _btrfs_profile_configs
>  }
>  
> -- 
> 2.4.6
> 
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Sterba July 24, 2015, 12:09 p.m. UTC | #2
On Thu, Jul 23, 2015 at 01:51:50PM -0700, Omar Sandoval wrote:
> +			# We can't do replace with these profiles because they
> +			# imply only one device ($SCRATCH_DEV), and we need to
> +			# keep $SCRATCH_DEV around for _scratch_mount
> +			# and _check_scratch_fs.
> +			local unsupported=(
> +				"single"
> +				"dup"

DUP does imply single device, but why does 'single' ?
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Omar Sandoval July 24, 2015, 5:50 p.m. UTC | #3
On Fri, Jul 24, 2015 at 02:09:46PM +0200, David Sterba wrote:
> On Thu, Jul 23, 2015 at 01:51:50PM -0700, Omar Sandoval wrote:
> > +			# We can't do replace with these profiles because they
> > +			# imply only one device ($SCRATCH_DEV), and we need to
> > +			# keep $SCRATCH_DEV around for _scratch_mount
> > +			# and _check_scratch_fs.
> > +			local unsupported=(
> > +				"single"
> > +				"dup"
> 
> DUP does imply single device, but why does 'single' ?

It does not, I apparently forgot that you could use single to
concatenate multiple devices. I'll fix that in v2.

Thanks for reviewing!
Brendan Hide July 27, 2015, 10:22 p.m. UTC | #4
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2015/07/24 07:50 PM, Omar Sandoval wrote:
> On Fri, Jul 24, 2015 at 02:09:46PM +0200, David Sterba wrote:
>> On Thu, Jul 23, 2015 at 01:51:50PM -0700, Omar Sandoval wrote:
>>> +			# We can't do replace with these profiles because they +
>>> # imply only one device ($SCRATCH_DEV), and we need to +			#
>>> keep $SCRATCH_DEV around for _scratch_mount +			# and
>>> _check_scratch_fs. +			local unsupported=( +				"single" +
>>> "dup"
>> 
>> DUP does imply single device, but why does 'single' ?
> 
> It does not, I apparently forgot that you could use single to 
> concatenate multiple devices. I'll fix that in v2.
> 
> Thanks for reviewing!
> 
Late to the party. DUP *implies* single device but there are cases
where dup is used on a multi-device fs. Even if the use-cases aren't
good or intended to be long-term, they are still valid, right?


- --
Hugo Mills July 27, 2015, 10:44 p.m. UTC | #5
On Tue, Jul 28, 2015 at 12:22:56AM +0200, Brendan Hide wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 2015/07/24 07:50 PM, Omar Sandoval wrote:
> > On Fri, Jul 24, 2015 at 02:09:46PM +0200, David Sterba wrote:
> >> On Thu, Jul 23, 2015 at 01:51:50PM -0700, Omar Sandoval wrote:
> >>> +			# We can't do replace with these profiles because they +
> >>> # imply only one device ($SCRATCH_DEV), and we need to +			#
> >>> keep $SCRATCH_DEV around for _scratch_mount +			# and
> >>> _check_scratch_fs. +			local unsupported=( +				"single" +
> >>> "dup"
> >> 
> >> DUP does imply single device, but why does 'single' ?
> > 
> > It does not, I apparently forgot that you could use single to 
> > concatenate multiple devices. I'll fix that in v2.
> > 
> > Thanks for reviewing!
> > 
> Late to the party. DUP *implies* single device but there are cases
> where dup is used on a multi-device fs. Even if the use-cases aren't
> good or intended to be long-term, they are still valid, right?

   Yes, but bear in mind that any new (presumably metadata) chunks
allocated on the multi-device FS will end up as RAID-1, because of the
automatic upgrade from DUP to RAID-1. Any balance will also do the
upgrade.

   Hugo.

PS. Can we get rid of that upgrade? It's a pain the arse in so many
ways.

> - -- 
> __________
> Brendan Hide
> http://swiftspirit.co.za/
> http://www.webafrica.co.za/?AFF1E97
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.22 (MingW32)
> 
> iQEcBAEBAgAGBQJVtq9AAAoJEE+uni74c4qN5eYIAJAGznsi3RD1tchbSLwhMXJk
> bJJ4ORB9taLXHykSfYTsHIaUoVpcVR6tT/I1jz5070DY3mKkQ16a8nwtSxPba4Lv
> QiS8YRegFiHMYzZbH1T7Tnm6R9g/RZsaU7GS3JhP9HUYG7hIWGRRuoiOjYn/hoLw
> uMXuIFOkPKGYDgyAhDIp3KDYlBjMHT6Oun7CcpvTjXiOnzJFFp3MSt3b6mmmdMVV
> YKWpWyKVh7qlENEoqKb4exqr1WGYKU+kBLXRs4wdm3xb66EcWYs0Er1u6v+K1trx
> nryFrfUxYtMJsSuR9ZJm88DOsXKAuX1LEdRKVOlq7krsIK8HlTizccMXAl10gKk=
> =ndkL
> -----END PGP SIGNATURE-----
David Sterba July 28, 2015, 4:52 p.m. UTC | #6
On Tue, Jul 28, 2015 at 12:22:56AM +0200, Brendan Hide wrote:
> > It does not, I apparently forgot that you could use single to 
> > concatenate multiple devices. I'll fix that in v2.
> > 
> > Thanks for reviewing!
> > 
> Late to the party. DUP *implies* single device but there are cases
> where dup is used on a multi-device fs. Even if the use-cases aren't
> good or intended to be long-term, they are still valid, right?

You're right, DUP is reported by 'fi df' after 2nd device is added and
this state (even if it's temporary) has to be taken into account.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Omar Sandoval July 28, 2015, 5:41 p.m. UTC | #7
On Tue, Jul 28, 2015 at 06:52:07PM +0200, David Sterba wrote:
> On Tue, Jul 28, 2015 at 12:22:56AM +0200, Brendan Hide wrote:
> > > It does not, I apparently forgot that you could use single to 
> > > concatenate multiple devices. I'll fix that in v2.
> > > 
> > > Thanks for reviewing!
> > > 
> > Late to the party. DUP *implies* single device but there are cases
> > where dup is used on a multi-device fs. Even if the use-cases aren't
> > good or intended to be long-term, they are still valid, right?
> 
> You're right, DUP is reported by 'fi df' after 2nd device is added and
> this state (even if it's temporary) has to be taken into account.

I wouldn't read too much into this, _btrfs_get_profile_configs is just
for tests that mkfs a bunch of new filesystems for testing, and in that
case, DUP is invalid for multiple devices. It's not meant to be an
authoritative source of truth regarding what metadata/data profiles
could be legal.

Thanks,
diff mbox

Patch

diff --git a/common/rc b/common/rc
index 610045eab304..3e6fdb6ebcfa 100644
--- a/common/rc
+++ b/common/rc
@@ -2748,60 +2748,62 @@  _btrfs_get_profile_configs()
 		return
 	fi
 
-	# no user specified btrfs profile configs, export the default configs
 	if [ -z "$BTRFS_PROFILE_CONFIGS" ]; then
-		# default configs
-		_btrfs_profile_configs=(
-			"-m single -d single"
-			"-m dup -d single"
-			"-m raid0 -d raid0"
-			"-m raid1 -d raid0"
-			"-m raid1 -d raid1"
-			"-m raid10 -d raid10"
-			"-m raid5 -d raid5"
-			"-m raid6 -d raid6"
+		# Default configurations to test.
+		local configs=(
+			"single:single"
+			"dup:single"
+			"raid0:raid0"
+			"raid1:raid0"
+			"raid1:raid1"
+			"raid10:raid10"
+			"raid5:raid5"
+			"raid6:raid6"
 		)
+	else
+		# User-provided configurations.
+		local configs=(${BTRFS_PROFILE_CONFIGS[@]})
+	fi
 
-		# remove dup/raid5/raid6 profiles if we're doing device replace
-		# dup profile indicates only one device being used (SCRATCH_DEV),
-		# but we don't want to replace SCRATCH_DEV, which will be used in
-		# _scratch_mount/_check_scratch_fs etc.
-		# and raid5/raid6 doesn't support replace yet
+	_btrfs_profile_configs=()
+	for cfg in "${configs[@]}"; do
+		local supported=true
+		local profiles=(${cfg/:/ })
 		if [ "$1" == "replace" ]; then
-			_btrfs_profile_configs=(
-				"-m single -d single"
-				"-m raid0 -d raid0"
-				"-m raid1 -d raid0"
-				"-m raid1 -d raid1"
-				"-m raid10 -d raid10"
-				# add these back when raid5/6 is working with replace
-				#"-m raid5 -d raid5"
-				#"-m raid6 -d raid6"
+			# We can't do replace with these profiles because they
+			# imply only one device ($SCRATCH_DEV), and we need to
+			# keep $SCRATCH_DEV around for _scratch_mount
+			# and _check_scratch_fs.
+			local unsupported=(
+				"single"
+				"dup"
 			)
+		elif [ "$1" == "replace-missing" ]; then
+			# We can't replace missing devices with these profiles
+			# because there isn't enough redundancy.
+			local unsupported=(
+				"single"
+				"dup"
+				"raid0"
+			)
+		else
+			local unsupported=()
 		fi
-		export _btrfs_profile_configs
-		return
-	fi
-
-	# parse user specified btrfs profile configs
-	local i=0
-	local cfg=""
-	for cfg in $BTRFS_PROFILE_CONFIGS; do
-		# turn "metadata:data" format to "-m metadata -d data"
-		# and assign it to _btrfs_profile_configs array
-		cfg=`echo "$cfg" | sed -e 's/^/-m /' -e 's/:/ -d /'`
-		_btrfs_profile_configs[$i]="$cfg"
-		let i=i+1
-	done
-
-	if [ "$1" == "replace" ]; then
-		if echo ${_btrfs_profile_configs[*]} | grep -q raid[56]; then
-			_notrun "RAID5/6 doesn't support btrfs device replace yet"
-		fi
-		if echo ${_btrfs_profile_configs[*]} | grep -q dup; then
-			_notrun "Do not set dup profile in btrfs device replace test"
+		for unsupp in "${unsupported[@]}"; do
+			if [ "${profiles[0]}" == "$unsupp" -o "${profiles[1]}" == "$unsupp" ]; then
+			     if [ -z "$BTRFS_PROFILE_CONFIGS" ]; then
+				     # For the default config, just omit it.
+				     supported=false
+			     else
+				     # For user-provided config, don't run the test.
+				     _notrun "Profile $unsupp not supported for $1"
+			     fi
+			fi
+		done
+		if "$supported"; then
+			_btrfs_profile_configs+=("-m ${profiles[0]} -d ${profiles[1]}")
 		fi
-	fi
+	done
 	export _btrfs_profile_configs
 }