Filesystem extension issues in HP-UX

[gnana prakasam]

I am getting the below error when extending the filesystem on HP-UX .

UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgtest/rlvtest1 failed with errno 16.

What could be the issue here  ?

[Steven Schweda]

> I am getting the below error when extending the filesystem [...]

   As usual, showing an actual command can be more helpful than a vague
description.

> [...] on HP-UX .

      uname -a

> [...] errno 16.
>
> What could be the issue here  ?

   I know nothing, but errno 16 is EBUSY, "Mount device busy".

[Patrick Wallek]

If you are trying to use the 'extendfs' command on the LV, then you must first unmount the LV.

[Bill Hassell]

If the mountpoint you are extending has been exported using NFS or SAMBA, it will be marked as busy. The mountpoint must un-exported before you can extend it.

[gnana prakasam]

Hi All,

Here the command used to extend the filesystem and failure message.

root@testserv1[/] # lvextend -L 35840M /dev/vgtest/lvtest1
Logical volume "/dev/vgtest/lvtest1" has been successfully extended.
Volume Group configuration for /dev/vgtest has been saved in /etc/lvmconf/vgtest.conf
root@testserv1[/] # fsadm -F vxfs -b 35840M /test1
UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgtest/rlvtest1 failed with errno 16
root@testserv1[/] #

This Filesystem is not shared as NFS & Samba. I am using lvextend & fsadm to extend the filesystem.

HP-UX Version is B.11.31.

This filesystem snaped with other filesystem. Please let me know the procedure to extend .

root@testserv1[/] # mount -v |grep -i test1
/dev/vgtest/lvtest1 on /test1 type vxfs ioerror=mwdisable,largefiles,delaylog,dev=80000003 on Sat Feb 6 13:08:11 2016
/dev/vgtest/backuptest1 on /backuptest1 type vxfs ro,ioerror=mwdisable,largefiles,snapof=/dev/vgtest/lvtest1,snapsize=13000000,dev=80000006 on Sun Feb 14 00:30:05 2016
root@testserv1[/] #

Aplogies for the late reply

[rescuehp17]

What is the solution for this error, I notice someone faced the same issue.

(root) @/> fsadm -F vxfs -b 330227968 /oracle
UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgoracle/rlvoracle failed with errno 16
(root) @/>

Defragmented, still no help..

Thank you in advance!

[Torsten.]

The error means device busy. Sure you have OnlineJFS?

[rescuehp17]

Yes, have online JFS as I was able to do 5 filesystems and other 5 filesystems failed with the same error mentioned above. 

[rescuehp17]

(root) @/> swlist -l product | grep -i online
AONLINEJFS B.05.00.02 OnlineJFS Integration Product
OnlineJFS01 5.0.1 Online features of the VxFS File System
(root)@/>

All the ten file systems are from one volume group, and out of 10 successfully I was able to extend FS size on five and the other five failed with error 16. 

Help is appreciated.

[Bill Hassell]

Your command:

fsadm -F vxfs -b 330227968 /oracle

is to resize the /oracle mountpoint to 330 million sectors. Is that correct?
Or did you mean 330 GB? If so, you'll need to use a suffix such as -b 330G.

Also, is /oracle a top level filesystem or does it contain other lvols or NFS mounts?
      hint:    bdf   /oracle
That will show the source lvol for the mountpoint.

And did you first increase the source lvol size first using lvextend?

[rescuehp17]

It was 220Gig before and I extended 100Gig to the lvm and so extending to 330Gig with fsadm

[Bill Hassell]

Is there a reason that the rather unusual number 330227968  is being used?
fsadm has been historically unstable due to the "sectors" terminology.
Prior to HP-UX 11.31, the man page was very bad as it failed to document the secret option modifiers.
From 11.23:

 -b newsize  Resize the file system to newsize sectors.  
                     -b cannot  be used with the -cdDeEo options.

And from 11.11:

 -b newsize  Resize the file system to newsize sectors.

Is a sector 256 bytes? 512 bytes? 1024 bytes? 2048 bytes? How about 4096 byes?
The term sector is commonly used with disk hardware and is independent from commands used in HP-UX.
There are several man pages that mention "sector" but the exact value is not an stable term.
Here are the 1m man pages that mention sector:

diskinfo.1m
disksecn.1m
extendfs_hfs.1m
extendfs_vxfs.1m
format.1m
fsadm_vxfs.1m
mkfs_hfs.1m
mkfs_vxfs.1m
mount_vxfs.1m
ncheck_hfs.1m
ncheck_vxfs.1m
newfs_hfs.1m
newfs_vxfs.1m
umount_vxfs.1m
vxassist.1m
vxdisk.1m
vxdisksetup.1m
vxdiskunsetup.1m
vxintro.1m
vxmake.1m
vxprint.1m
vxrecover.1m
vxtunefs.1m
vxvol.1m

S the use of sectors as a metric is too unreliable for an exact meaning.
The fsadm man page never defines what a sector is. 

Always use fsadm -b with a dependable metric as in: -b 320000M or -b 320G

[rescuehp17]

That didn't work either.

[Bill Hassell]

Without the actual error message, I can only guess that it was the same message.
Post the output from vgdisplay for that lvol. Is this lvol perhaps a distributed layout? What does the command: fstyp /dev/vg00/xyz (where xyz is the knoll name) report?

[rescuehp17]

root) @/> fsadm -F vxfs -b 330G /oracle

UX:vxfs fsadm: ERROR: V-3-25811: cannot expand /dev/vgoracle/rlvoracle more than size of the underlying device - 332660736 sectors

(root) @/> fsadm -F vxfs -b 332660736 /oracle

UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgoracle/rlvoracle failed with errno 16

(root) @/> fsadm -F vxfs -b 330000000 /oracle

UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgoracle/rlvoracle failed with errno 16

(root) @/tmp> fsadm -F vxfs -b 320G /oracle
UX:vxfs fsadm: ERROR: V-3-25811: cannot expand /dev/vgoracle/rlvoracle more than size of the underlying device - 332660736 sectors
(root) @/tmp>

---------------- lvdisplay -----------------------------

--- Logical volumes ---
LV Name /dev/vgoracle/lvoracle
VG Name /dev/vgoracle
LV Permission read/write
LV Status available/syncd
Mirror copies 0
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 324864
Current LE 5076
Allocated PE 5076
Stripes 0
Stripe Size (Kbytes) 0
Bad block on
Allocation strict
IO Timeout (Seconds) default

--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/disk/disk109 1520 1520
/dev/disk/disk138 640 640
/dev/disk/disk284 1040 1040
/dev/disk/disk302 313 313
/dev/disk/disk355 253 253
/dev/disk/disk360 1310 1310

--- Logical extents ---
LE PV1 PE1 Status

(root)@/tmp> fstyp /dev/vgoracle/lvoracle
vxfs
(root)@/tmp>

[Bill Hassell]

From lvdisplay: LV Size (Mbytes) 324864

The size error is a possible clue. How about using -b 310G or -b 300G

Also, post the vgdisplay summary too.

[rescuehp17]

(root) @/tmp> fsadm -F vxfs -b 310G /oracle
UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgoracle/rlvoracle failed with errno 16
(root) @/tmp>

(root) @/tmp> fsadm -F vxfs -b 300G /oracle
UX:vxfs fsadm: ERROR: V-3-20340: attempt to resize /dev/vgoracle/rlvoracle failed with errno 16
(root) @/tmp>

So many lvm, so just displaying the below:

vgdisplay -v vgoracle | more

--- Volume groups ---
VG Name /dev/vgoracle
VG Write Access read/write
VG Status available
Max LV 255
Cur LV 94
Open LV 94
Max PV 64
Cur PV 33
Act PV 33
Max PE per PV 65535
VGDA 66
PE Size (Mbytes) 64
Total PE 498047
Alloc PE 497702
Free PE 345
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0
VG Version 1.0
VG Max Size 262140g
VG Max Extents 4194240

LV Name /dev/vgoracle/lvoracle
LV Status available/syncd
LV Size (Mbytes) 324864
Current LE 5076
Allocated PE 5076
Used PV 6

[Torsten.]

how about the output of

# mount

when you grow the diskspace, did you grow the vg and lvol too?

[rescuehp17]

Ofcourse, I exetended volume group and then logical volume and then fsadm, thats where it is failing.

(root) @/tmp> mount| grep lvoracle
/oracle on /dev/vgoracle/lvoracle ioerror=mwdisable,largefiles,delaylog,dev=403a0046 on Wed Oct 7 20:11:04 2015

[Torsten.]

Check for possible show stoppers:

• The file system has active mounted checkpoints
• File system has been flagged as requiring an FSCK in the superblock
• The file system is too busy for resize locks to be obtained

A suggested solution would be

unmount the file system, perform "fsck -o full", remount, and try again.

(found at https://support.symantec.com )

[rescuehp17]

Can't have downtime. So need to be done online. Defragmentation is done, still non't work.

[rescuehp17]

Any ideas, guys? Really appreciate and this project is pending.

[rescuehp17]

(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo10 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo30 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo50 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo70 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo80 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp>

[rescuehp17]

Still no luck to increase the file system size:

(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo10 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo30 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo50 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo70 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp> echo "8192B.p S" | /opt/VRTS/bin/fsdb /dev/vgtest/rlvbo80 | grep flags
flags 0 mod 0 clean 3c
(root) @/tmp>

[Torsten.]

I googled this:

https://sort.veritas.com/public/documents/sfha/6.0/aix/productguides/html/sfcfs_install/ch11s04s05.htm

A clean_value value of 0x5a indicates the file system is clean. A value of 0x3c indicates the file system is dirty.