08-08-2011 08:16 PM
Dear all,
I have the resize the FS in that I made a mistake. The below are the commmands I have executed
fstyp /dev/vx/dsk/ua1195dg/duo1290p-oradata-fs
ufs
vxresize -g ua1195dg -F vxfs -x duo1290p-oradata-fs +100g // I have mistakenly resized the FS as ( -F vxfs )
The FS got increased without any error and data started storing.
1) Hence in this case, I have to do anything ?
2) It will affect the FS while rebooting ?
Any suggestion please
Solved! Go to Solution.
08-09-2011 08:14 AM
Hi
Not seen that done before, but it appears to have actually used the UFS grow operation, even though you used the vxfs switch. I ran a trivial similar test and it also appears to have grown the FS. I also ran a truss to look for the the mkfs commands(notice its using ufs paths):
boo# vxassist -g testdg make tony1 100m
boo# newfs /dev/vx/rdsk/testdg/tony1
boo# mount -Fufs /dev/vx/dsk/testdg/tony1 /tony1
boo# df -k /tony1
Filesystem kbytes used avail capacity Mounted on
/dev/vx/dsk/testdg/tony1
95551 1041 84955 2% /tony1
boo# truss -failed -o /tmp/truss vxresize -g testdg -Fvxfs -x tony1 +100m
boo# df -k /tony1
Filesystem kbytes used avail capacity Mounted on
/dev/vx/dsk/testdg/tony1
192095 1041 181499 1% /tony1
boo# grep mkfs /tmp/truss
26310/1: /usr/lib/fs/ufs/mkfs -M "/tony1" "/dev/vx/rdsk/testdg/tony1" "409600" 2>&1
26312/1: 0.8867 execve("/usr/lib/fs/ufs/mkfs", 0x0003A7B4, 0x0003A7D0) argc = 5
26312/1: argv: /usr/lib/fs/ufs/mkfs -M /tony1 /dev/vx/rdsk/testdg/tony1
26312/1: 0.8880 resolvepath("/usr/lib/fs/ufs/mkfs", "/usr/lib/fs/ufs/mkfs", 1023) = 20
26312/1: 0.8881 stat64("/usr/lib/fs/ufs/mkfs", 0xFFBFF8F0) = 0
As a precaution you may want to take a backup of the file system
08-09-2011 01:51 AM
Hope you get a resolution here.
08-09-2011 02:49 AM
vxresize runs a volume resize and then a vxfs resize. I would have thought that the volume resize might have worked, but the filesystem resize would not - are you sure that the filesystem was increased - i.e does "df" show increased size?
Mike
08-09-2011 06:40 AM
Yes, the FS got increased without any error and data started storing.
1) Hence in this case, I have to do anything ?
2) It will affect the FS while rebooting ?
Any suggestions..
08-09-2011 08:14 AM
Hi
Not seen that done before, but it appears to have actually used the UFS grow operation, even though you used the vxfs switch. I ran a trivial similar test and it also appears to have grown the FS. I also ran a truss to look for the the mkfs commands(notice its using ufs paths):
boo# vxassist -g testdg make tony1 100m
boo# newfs /dev/vx/rdsk/testdg/tony1
boo# mount -Fufs /dev/vx/dsk/testdg/tony1 /tony1
boo# df -k /tony1
Filesystem kbytes used avail capacity Mounted on
/dev/vx/dsk/testdg/tony1
95551 1041 84955 2% /tony1
boo# truss -failed -o /tmp/truss vxresize -g testdg -Fvxfs -x tony1 +100m
boo# df -k /tony1
Filesystem kbytes used avail capacity Mounted on
/dev/vx/dsk/testdg/tony1
192095 1041 181499 1% /tony1
boo# grep mkfs /tmp/truss
26310/1: /usr/lib/fs/ufs/mkfs -M "/tony1" "/dev/vx/rdsk/testdg/tony1" "409600" 2>&1
26312/1: 0.8867 execve("/usr/lib/fs/ufs/mkfs", 0x0003A7B4, 0x0003A7D0) argc = 5
26312/1: argv: /usr/lib/fs/ufs/mkfs -M /tony1 /dev/vx/rdsk/testdg/tony1
26312/1: 0.8880 resolvepath("/usr/lib/fs/ufs/mkfs", "/usr/lib/fs/ufs/mkfs", 1023) = 20
26312/1: 0.8881 stat64("/usr/lib/fs/ufs/mkfs", 0xFFBFF8F0) = 0
As a precaution you may want to take a backup of the file system
08-12-2011 10:42 PM
Hi Tony,
The way you this found the result is very fantastic... Thank you very much :)
And I am having some doubt,
1) For this vxresize whether we can mkfs ?. If yes whats the command to resize using mkfs
2) In our environment, fstyp having ufs and vxfs. whether I need to use this -F vxfs / ufs ? or not this switch is not required ?. Which is best one
Thanks in advance,
Regards,
Senthilsam
08-12-2011 10:45 PM
Dear team,
While checking mutlipath, I am getting below error message as vxdmpadm: No such device
[root@up81075b GLOBAL] /dev/rdsk # vxdmpadm listctlr all
CTLR-NAME ENCLR-TYPE STATE ENCLR-NAME
=====================================================
c2 TagmaStore-USP ENABLED tagmastore-usp1
c4 TagmaStore-USP ENABLED tagmastore-usp1
c3 TagmaStore-USP ENABLED tagmastore-usp1
c1 TagmaStore-USP ENABLED tagmastore-usp1
c0 Disk ENABLED disk
[root@up81075b GLOBAL] /dev/rdsk # vxdmpadm getdmpnode enclosure=tagmastore-usp1
vxdmpadm: No such device
can you please confirm what causes this message. How to resolve it ?
Thanks in advance
Regards,
senthilsam
08-15-2011 03:25 AM
As this is a different problem it should be posted as a separate thread - appears to have been done here:
https://www-secure.symantec.com/connect/forums/multipath-vxdmpadm-no-such-device
08-15-2011 03:39 AM
1. If you need to resize a vxfs filesystem separately to the volume, you can use fsadm. You need to ensure the volume has been grown to the required size prior to resizing the volume. (NOTE: vxresize does both the volume and the filesystem in one step/command, so it is recommended to use vxresize rather than resizing components separately)
See: Veritas File System 5.1 Administrator's Guide (Solaris) - Appendix A. Quick Reference - Resizing a file system
https://sort.symantec.com/public/documents/sf/5.1/solaris/html/vxfs_admin/apas09.htm
See also the fsadm manpage (eg: for VxFS 5.1 Solaris):
https://sort.symantec.com/public/documents/sf/5.1/solaris/manpages/file_system/man/html/man1m/fsadm_vxfs.1m.html
Note the new size has to be specified in sectors (and is the total new size, not the amount you want to increase by). You still need to ensure you are running this on the correct filesystem type (ie: if you run fsadm on a ufs filesystem by mistake, you may encounter issues)
2. In future to avoid any potential issues, determine the filesystem type (ufs or vxfs) prior to running the vxresize, then run vxresize with the correct filesytem type (-F vxfs or -F ufs as appropriate) to avoid any confusion. Although in this case mistakenly specifying -F vxfs did not affect the ufs fileystem as it appears it recognised it was a ufs filesystem and used the correct commands accordingly, the reason for specifying the filesystem type is to ensure the correct commands are used to grow the filesystem, so you should be checking the filesystem type beforehand rather than relying on the command to check this for you!