11-06-2007 11:04 AM
Here is a good one.
Today I had to add some space to the volume ( /app ) on a production server which is in a cluster with another one. To be safe, I decided to do this first on a test server ( did I mantian this was the first time I am doing this? ).
Production server is running HP-UX 11.11 and test server is an older N-class with HP-UX 10.0
To acomplish this, I ran this commands:
fuser -cu /app ( i killed 4 running processes )
lvextend -L 1024 /dev/vg00/lvol8 /dev/dsk/c2t0d0 /dev/dsk/c1t0d0
extendfs -F vxfs /dev/vg00/lvol8
To my supprise, evrything worked just fine. I did the same thing on my production server. No problem.
However, after I rebooted the test server, I can not get it to start again.
I am getting error "Legacy PA HEX Chasis Code".
I did not reboot the production server so that one is ok. My question is, what have I done wrong? Am I going to have the same problem with production server after the reboot?
One thing that I am not sure about, is that I know disks were mirrored and I did not brake the mirror before extending the volume. Could this be a problem?
Any suggestion is wellcome.
Solved! Go to Solution.
11-06-2007 12:03 PM
Since the mirroring is done on logical volume level so once the LV is mirrored and then if you extend it , that will get exteded across mirrors provided there is enough space.
11-06-2007 04:06 PM
As said the error during boot is surely not related to the activity of adding space to the lvol.its related to something else
Try the following
1.Shutdown the test box and disconnect any sotrage connected to it.
2. Boot up the test server without any io connections externally.
3.If the system boots fine then insert the external cables one by one and repeat the shutdown and powerup procedure.
4. Any issues Check for the lastest HPMC /var/tombstones/ts99 timestamp. Check the syslog and eventlog for errors and the root cause can be traced.
11-07-2007 06:17 AM
A message like
alloc_pdc_pages: Relocating PDC from 0xffff800000 to 0x5faab000
is also a normal part of HP-UX boot.
Some server models will output this message and then spend a *lot* of time with no visible output. This is normal, although it can be hard on the nerves of a new sysadmin.
A full reboot of a rp5450 (also known as L2000) can take about 10 minutes when the system has a lot of memory installed. As the N-class of servers uses the same generation of technology as the L-class but has double the amount of CPUs and memory, I'd expect it to take even longer to boot.
The full list of chassis codes for L and N class servers seems to be HP internal information, but www.openpa.net website has a list of K class server chassis codes. The chassis codes of newer servers seem to use a similar structure in their chassis codes, so you can look at the first one or two digits of each code and find out the general meaning of the codes using the K class code listing.
Unfortunately the www.openpa.net site seems to be *very* slow at the moment, so I cannot provide an exact link for you.
Anyway, many of the chassis codes just document how far the selftest or boot has progressed. When HP-UX has started up, codes of the form FxxF actually indicate the number of active CPUs and the current load level of the machine.