Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations gkittelson on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

mirrorvg testing 1

Status
Not open for further replies.

promotor

Technical User
Oct 14, 2001
35
PH
how can i test if the mirroring of two internal disks (hdisk0 and hdisk1) work?
can i just remove the definition of let's say hdisk0 then reboot the system and see if it boots to hdisk1?
 
it is a bad idea to hold swap and jfslog on mirror - you should remove disk not only definition, and recovering is tricky anyway
 
Hi,

I've been dealing with this problem couple of weeks ago (see older threads).

The best solution was to phisically remove one of the disks.

Other dirty tricks (ODM hacking,etc) are NOT recommended,I had bad results and had to reinstall the whole system from scratch. "Long live king Moshiach !"
h
 
change the bootlist so that system boots form hdisk1 rather than hdisk0. So you can have bootlist as hdisk1 cd tape
 
As for "it is a bad idea to hold swap and jfslog on mirror ", this is absurd; when you mirror your volume group all lv's are mirrored, paging, jfslog, lv's, the only lv not mirrored is the sysdump device.
 
Mirroring paging and jfslog is OK, dump device not OK:

>lsvg -l rootvg
rootvg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
hd6 paging 200 400 2 open/syncd N/A
hd5 boot 1 2 2 closed/syncd N/A
hd8 jfslog 1 2 2 open/syncd N/A
hd4 jfs 30 60 2 open/syncd /
hd2 jfs 192 384 2 open/syncd /usr
hd9var jfs 13 26 2 open/syncd /var
hd3 jfs 11 22 2 open/syncd /tmp
hd1 jfs 13 26 2 open/syncd /home
dumplv0 sysdump 25 25 1 open/syncd N/A
dumplv1 sysdump 20 20 1 open/syncd N/A

As for verifying mirror, syncd is good, stale is bad. IBM Certified -- AIX 4.3 Obfuscation
 
In agreement with everyone else.

If you don't mirror your jfslog and paging space, you might not survive a disk failure. If the disk with those fails, the running system can no longer write to any rootvg filesystems, which is bad, but before that can happen, the kernel can't access paging space, which will crash your systems pretty much immediately.

Booting into a rootvg in a missing-drive config without those LVs having been mirrored will also be impossible - the filesystems cannot be mounted without a jfslog. I've never tried bringing a system up without any paging space, but I have a suspicion AIX won't come up - it'll probably stop with C51 on the front panel display.
 
Just FYI, AIX actually will boot without active paging space, but of course you need adequate RAM to do this. I found this out during some upgrades from 4.3.2 to 4.3.3 on F50 boxen when some sort of bug deactivated hd6 until I patched to ML-09 and rebooted. Second upgrade, I defined a temporary pagingvg so it would IPL in less than 45 minutes. IBM Certified -- AIX 4.3 Obfuscation
 
I'm running the following command in a script that performs a disk-to-disk backup;

dd if=/dev/hdiskx of=/dev/hdiskX bs=1024k

The problems is the disks must be the same size. You cannot go from say a 4.5 GB to a 9GB drive. Is there a
work around for this ?
 
Off the top of my head, duplicating an hdisk like that will also duplicate the pvid which can lead to confusing the ODM.

What exactly do you want to get done?
 
I would say the workaround is to use the LVM to mirror. It can even do a 3-way mirror if you want and is pretty transparent, meaning you don't have to run a cron job (or what have you) and therefore have the potential to lose a recent portion of your data. Plus someone already wrote it so why not use it?

I am assuming you are just trying to mirror your disks let me know if I missed your intention. IBM Certified -- AIX 4.3 Obfuscation
 
Status
Not open for further replies.

Similar threads

  • Locked
  • Question
Replies
8
Views
125
Replies
1
Views
49
Replies
11
Views
196
Replies
3
Views
110

Part and Inventory Search

Sponsor

Back
Top