Hi. I have a volume group (let's say) vg_data. It consists from /dev/sdd5 sdd6 sdd7 I added sdc5
Now I want to remove (free) sdd7 and you is to for RAID partition.
What are the commands (ordered) I need to perform? I failed to find clear howto.
vg-data has only one partition, total size is over 1TB, free space is about 500GB so there is a plenty or room.
----- Original Message ----- | Hi. I have a volume group (let's say) vg_data. | It consists from /dev/sdd5 | sdd6 | sdd7 | I added sdc5 | | Now I want to remove (free) sdd7 and you is to for RAID partition. | | What are the commands (ordered) I need to perform? I failed to find | clear howto. | | | vg-data has only one partition, total size is over 1TB, free space is | about 500GB so there is a plenty or room.
did you pvmove the extents that are on the drive you want to remove?
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Ljubomir Ljubojevic Sent: Tuesday, June 24, 2014 11:37 AM To: CentOS mailing list Subject: [CentOS] How to remove LVM Physical Volume from Volume Group?
Hi. I have a volume group (let's say) vg_data. It consists from /dev/sdd5 sdd6 sdd7 I added sdc5
Now I want to remove (free) sdd7 and you is to for RAID partition.
What are the commands (ordered) I need to perform? I failed to find clear howto.
vg-data has only one partition, total size is over 1TB, free space is about 500GB so there is a plenty or room.
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/...
-- Eero
The doc is recommend you do pvremove first and then run vgreduce. But you will get error something like "...you need to remove from volume group first before delete the physical volume" (the message makes sense for me. Need to removing from volume group first then removing the physical one)
While ago I run "vgrreduce" first and then do "pvremove". It worked fine for me.
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Eero Volotinen Sent: Tuesday, June 24, 2014 12:45 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/...
-- Eero _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Tue, Jun 24, 2014 at 3:56 PM, Zhang, Jonathan zhangj@evergreen.edu wrote:
The doc is recommend you do pvremove first and then run vgreduce. But you will get error something like "...you need to remove from volume group first before delete the physical volume" (the message makes sense for me. Need to removing from volume group first then removing the physical one)
While ago I run "vgrreduce" first and then do "pvremove". It worked fine for me.
I myself always thought you would pvmove, then vgreduce, and then pvremove. Unless you also had to resize the lv first
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Eero Volotinen Sent: Tuesday, June 24, 2014 12:45 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/...
-- Eero _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
If you need to move data out from sdd7 to new location. You need to pvmove /dev/new_disc first Then vgreduce and pvremove
However, before do pvmove you need pvcreate /dev/new_disc vgextend vg_data/new_disc pvmove /dev/new_disc
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Mauricio Tavares Sent: Tuesday, June 24, 2014 1:00 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
On Tue, Jun 24, 2014 at 3:56 PM, Zhang, Jonathan zhangj@evergreen.edu wrote:
The doc is recommend you do pvremove first and then run vgreduce. But you will get error something like "...you need to remove from volume group first before delete the physical volume" (the message makes sense for me. Need to removing from volume group first then removing the physical one)
While ago I run "vgrreduce" first and then do "pvremove". It worked fine for me.
I myself always thought you would pvmove, then vgreduce, and then pvremove. Unless you also had to resize the lv first
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Eero Volotinen Sent: Tuesday, June 24, 2014 12:45 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_ Linux/6/html/Logical_Volume_Manager_Administration/disk_remove_ex.html
-- Eero _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
_______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 06/24/2014 09:59 PM, Mauricio Tavares wrote:
On Tue, Jun 24, 2014 at 3:56 PM, Zhang, Jonathan zhangj@evergreen.edu wrote:
The doc is recommend you do pvremove first and then run vgreduce. But you will get error something like "...you need to remove from volume group first before delete the physical volume" (the message makes sense for me. Need to removing from volume group first then removing the physical one)
While ago I run "vgrreduce" first and then do "pvremove". It worked fine for me.
I myself always thought you would pvmove, then vgreduce, andthen pvremove. Unless you also had to resize the lv first
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Eero Volotinen Sent: Tuesday, June 24, 2014 12:45 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/...
--
As like many time before, as soon as Isent the e-mail with a question, a lightbulb went off in my mind.
1. Unmount filesystem (I only have one in the group) on Logical Volume. 2. Resize Logical Volume since I stretched it before (did not know how all that goes). I resized it to well over the size of sdd7. I used LVM GUI for filesystem+LV resize in one go.
Now: 3. pvmove to move away all allocated extents 4. vgreduce
I hope those last too go without a problem. Will report back.
On 06/24/2014 10:11 PM, Ljubomir Ljubojevic wrote:
On 06/24/2014 09:59 PM, Mauricio Tavares wrote:
On Tue, Jun 24, 2014 at 3:56 PM, Zhang, Jonathan zhangj@evergreen.edu wrote:
The doc is recommend you do pvremove first and then run vgreduce. But you will get error something like "...you need to remove from volume group first before delete the physical volume" (the message makes sense for me. Need to removing from volume group first then removing the physical one)
While ago I run "vgrreduce" first and then do "pvremove". It worked fine for me.
I myself always thought you would pvmove, then vgreduce, andthen pvremove. Unless you also had to resize the lv first
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Eero Volotinen Sent: Tuesday, June 24, 2014 12:45 PM To: CentOS mailing list Subject: Re: [CentOS] How to remove LVM Physical Volume from Volume Group?
2014-06-24 22:31 GMT+03:00 Zhang, Jonathan zhangj@evergreen.edu:
Fidsk -l (you will see sdd5, sdd6 and sdd7) Pvdisplay (to make sure the Allocated PE on /dev/sdd7 Vgreduce vg_data /dev/sdd7 Pvremove /dev/sdd7
(before you run vgreduce you need to move all data from sdd7 to new dic
this is wrong way to do this. see the real docs at: https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/...
--
As like many time before, as soon as Isent the e-mail with a question, a lightbulb went off in my mind.
- Unmount filesystem (I only have one in the group) on Logical Volume.
- Resize Logical Volume since I stretched it before (did not know how
all that goes). I resized it to well over the size of sdd7. I used LVM GUI for filesystem+LV resize in one go.
Now: 3. pvmove to move away all allocated extents 4. vgreduce
I hope those last too go without a problem. Will report back.
It turned out that new partition, sdc5 is smaller then sdd7, so pvmove failed.
But then I checked the LVM GUI and solved the problem. I opened Physical volumes, chose sdd7 and clicked on "Migrate Extents....." which offered to redistribute all of the Extents automatically to the rest of the PV's.
At the moment I am waiting for migration to complete, but then removal of PV should be simple. I think I will use wonderful LVM GUI again, and I will not allow anyone to tell me how DE/GUI on a server is bad thing. ;)
On 06/24/2014 11:45 PM, Ljubomir Ljubojevic wrote:
It turned out that new partition, sdc5 is smaller then sdd7, so pvmove failed.
But then I checked the LVM GUI and solved the problem. I opened Physical volumes, chose sdd7 and clicked on "Migrate Extents....." which offered to redistribute all of the Extents automatically to the rest of the PV's.
At the moment I am waiting for migration to complete, but then removal of PV should be simple. I think I will use wonderful LVM GUI again, and I will not allow anyone to tell me how DE/GUI on a server is bad thing.;)
There are lots of rumors regarding the existence of a GUI. Sometimes it's a bit weird to use very low level tools when you have tools that will give you much more power in your hands.
But again some prefer to know what they do in the lowest levels and others just want to do the thing while they know it works and was designed to work. When the two collide ( works vs power) lots prefer to have only works (what seems to be a very bad assumption).
Many times you don't need GUI for servers to operate and sometimes the existence of a GUI slow the server down and lowering the performance of the system.
So taking couple things into account most admins agree that a GUI should not be needed when you know how to operate the machine using the CLI tools. If you feel the need to use GUI tools have it your way and have a nice smile to your CLI friends which struggle every once in a while how to do things in a way they prefer and note that not all admins are alike.
(I am still looking for a GUI utils to operate GlusterFS which seems to be there in RH storage systems)
Regards, Eliezer
[Top posting intentionally].
This e-mail triggers the Thunderbird bug, previously discussed,for me. The error message is: Bad key or directory name: "/desktop/gnome/url-handlers/GMT+03/command": `+' is an invalid character in key/directory names
and appears to be triggered by the + in the timezone.
On 06/24/2014 03:44 PM, Eero Volotinen wrote: ......
On 06/25/2014 10:30 AM, Lamar Owen wrote:
[Top posting intentionally].
This e-mail triggers the Thunderbird bug, previously discussed,for me. The error message is: Bad key or directory name: "/desktop/gnome/url-handlers/GMT+03/command": `+' is an invalid character in key/directory names
and appears to be triggered by the + in the timezone.
I've been curious about this; I'm using thunderbird but I don't see this error condition. perhaps the thunderbird I'm running from ubuntu just already has the patch.
zep wrote:
On 06/25/2014 10:30 AM, Lamar Owen wrote:
[Top posting intentionally].
This e-mail triggers the Thunderbird bug, previously discussed,for me. The error message is: Bad key or directory name: "/desktop/gnome/url-handlers/GMT+03/command": `+' is an invalid character in key/directory names
and appears to be triggered by the + in the timezone.
I've been curious about this; I'm using thunderbird but I don't see this error condition. perhaps the thunderbird I'm running from ubuntu just already has the patch.
It must. I've been annoyed for months, and I see it every day....
mark
On 06/26/2014 02:30 AM, Lamar Owen wrote:
[Top posting intentionally].
This e-mail triggers the Thunderbird bug, previously discussed,for me. The error message is: Bad key or directory name: "/desktop/gnome/url-handlers/GMT+03/command": `+' is an invalid character in key/directory names
+1 I get this on some messages - this one certainly - using latest CentOS thunderbird.
and appears to be triggered by the + in the timezone.
On 06/24/2014 03:44 PM, Eero Volotinen wrote: ...... _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-
Eero Volotinen -
Eliezer Croitoru -
James A. Peltier -
Lamar Owen -
Ljubomir Ljubojevic -
m.roth＠5-cent.us -
Mauricio Tavares -
Rob Kampen -
zep -
Zhang, Jonathan