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

Logical Volume problem in FASTt500 1

Status
Not open for further replies.

arvibm

Vendor
Feb 6, 2003
115
IN
I have four p-series server let's name them s1,s2,s3,s4 out of s1 and s3 shares the fastt500 storage through a san card.i have upgraded s3 server from aix 4.3.3 to aix 5.1(i used preservation install) but when i rebooted server after completion of installation the server has stopped showing the FASTt500 harddisks because during preservation install it overwrites the usr filesystem.i have reinstalled the san card driver on aix then it start showing harddisk of fastt500 but the problem is it pickups hdisk6 and hdisk7 from fastt500 but previously(before upgradation)it was hdisk2 and hdisk3 then i have to do lot of exercise to bring my volumegroup back using command exportvg,redfinevg -d vgname,importvg etc.then the vg is imported on hdisk7.but the actual problem comes when i rebooted s1 server the server is showing fastt500 disks but no vg mounted on it then i have imported the vg of s1 server on s3 server and on s3 server i can see the logical volume information of volume group of s1 server then i have exported the vg and on s1 server i have removed all the harddisk of fastt500(rmdev -dl <hdisk-name>) .then i run the command cfgmgr -i /dev/cd0 by inserting 1st installation cd of aix 4.3.3 and now it starts showing the harddisk with identifier name.then i imported the vg and i can see the logical volumes in volume groups but my problem is when i reboot the s1 server i have to do the same exercise again because the s1 server not automatically mounts the lv and vg on fastt500.while on s3 server(on which i have upgraded to aix 5.1) i don't have any problem s3 server is workin perfectly normal.On FASTt500 we have 16 disks two are hot swapable and out of 14 two arrays are defined assigned to s1 and s3 server.is there any solution to resolve the compatibility between 5.1 and 4.3.3??PLS HELP!!!!Thanx in advance...
 
I am having great difficulty in reading your question. Please use some more conventional form of punctuation in future.

It would be better to have done the upgrade as migration rather than preservation, but I think you know that now.

Is this supposed to be a concurrent VG (available to both systems simultaneously), or is it supposed to be a failover VG (known to both systems but with only one having access at any one time)?
 
Hi sectorseveng ,

Sorry for the inconvenience.I am not very sure about the FASTt500 as i am not the person who has initially installed and configured the fastt500.On FASTt500 we have two arrays each consists of 7 disks of 73.4GB.but i think it is failover vg because on s1 and s3 servers.
here is lspv output of s3 server:-
# lspv
hdisk0 000893da0174f887 rootvg
hdisk1 000893da811d5592 test
hdisk6 000893fad4aab9d6 None
hdisk7 000893da2f1433ff testvg
hdisk1 and hdisk 0 are internal harddisk of s3 server(aix 5.1) and hdisk6 and hdisk7 are harddisks of fastt500.

lspv output of s1 server:-
# lspv
hdisk0 000893fa65e36eae rootvg
hdisk1 000893fa2c5a8b4a rootvg
hdisk2 000893fad4aab9d6 dbvg
hdisk3 none None

hdisk0 and hdisk1 are internal harddisk of s1 server(aix 4.3.3)
hdisk2 and hdisk3 are FASTt500 disks.

i don't have any problem with s3 server the only problem is with s1 server.when i reboot the s1 server the logical volumes in volume group dbvg are not mounted though i have vg entries in /etc/filesystems.then i have to import the volume group of s1 server i.e. dbvg on s3 server and the export it again and then on s1 server i have remove the harddisks and reconfigure it using cfgmgr -i /dev/cd0.to assign the vg identifier i.d. i am using command chdev -l hdisk2 pv=yes.then the volume &quot;dbvg&quot; is imported and can be varied on.Then i can mount the volume group by mount all command.this exercise i have to do every time i restart s1 server.
i will be really grateful to you if u can help me to sort out this problem.
 
OK let me see if I understand this, you have 2 volume groups:

testvg on PVID 000893da2f1433ff which should be varied on by the s3 server.
dbvg on PVID 000893fad4aab9d6 which should be varied on by the s1 server.

Is the dbvg set to automatically varyon on the s1 server? What is the result of 'lsvg dbvg' when run on the s1 server?
 
output of lsvg dbvg on s1 server:-
# lsvg dbvg
VOLUME GROUP: dbvg VG IDENTIFIER: 000893fad4aafab0
VG STATE: active PP SIZE: 512 megabyte(s)
VG PERMISSION: read/write TOTAL PPs: 819 (419328 megabytes)
MAX LVs: 256 FREE PPs: 110 (56320 megabytes)
LVs: 15 USED PPs: 709 (363008 megabytes)
OPEN LVs: 15 QUORUM: 2
TOTAL PVs: 1 VG DESCRIPTORS: 2
STALE PVs: 0 STALE PPs: 0
ACTIVE PVs: 1 AUTO ON: yes
MAX PPs per PV: 1016 MAX PVs: 32

yes u r correct dbvg is the vg to be varied on s1 server and testvg is the vg to be varied on s3 server.hdisk7 on s3 server is same as hdisk3 on s1 server and hdisk6 on s3 server is same as hdisk2 on s1 server(hdisk identifiers are same).what do u mean by this question&quot;Is the dbvg set to automatically varyon on the s1 server?&quot;we have entries in /etc/filesystems for dbvg logical volumes.one thing more before upgrading the 5.1 on s3 server the lspv output is
#lspv

# lspv
hdisk0 000893da0174f887 rootvg
hdisk1 000893da811d5592 test
hdisk2 000893fad4aab9d6 None
hdisk3 000893da2f1433ff testvg

and after upgradation in place of hdisk2 and hdisk3 we have hdisk7 and hdisk6.
pls let me know if u need further details.waiting for ur response...

thanx and regards

arvind shrivastava
 
Arvind,

Having the filesystems set to automount in /etc/filesystems is not enough, the VG must be set to autovaryon in order to allow the logical volumes to be used. You can see this from the value of AUTO ON from the lsvg output, in your case this is set to yes so this is OK for s1.

If the VG is known to 2 systems it should only be set to auto varyon on one of them, this is because the first system to come up will acquire this volume group and prevent any other system from seeing it. ie, if dbvg is known to s1 & s3 and is set to autovaryon on both systems then if both are rebooted at the same time and s3 comes up first it will varyon dbvg and prevent s1 from seeing it, is this what is happening in your case?

I am also concerned that you have use the 'cfgmgr -i /dev/cd0' when reacquiring the disks on s1, you should only need to load things from the CD once not on every reboot.

What exactly is the status of s1 after a reboot but before you take these actions? After rebooting s1 what is the output of lspv? Are there any error messages in /var/adm/ras/bootlog? Are there any error messages in errpt?
 
Hi,

dbvg is known to s1 server and testvg is known to s3 server but as the fastt500 is shared between two servers i can import any of the two arrays defined on fastt500.
here is the output of /var/adm/ras/bootlog file:-
*************************************************************
ùóùô^T^G^Ë tory in the path name does not exist.
mount: 0506-324 Cannot mount /dev/lv01 on /usr/welcome: A file or directory in the path name does not exist.
mount: 0506-324 Cannot mount /dev/dump_lv on /dumpfs: The requested resource is busy.
mount: 0506-324 Cannot mount /dev/test_lv on /testfm: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/crp_lv on /crpfm: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/ora1_lv on /ora1: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/fslv00 on /d01: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/ora3_lv on /u3: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/fslv01 on /u4: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/loc_lv on /locfm: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/patch_lv on /patches: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/fslv02 on /visfm: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/testdata_lv on /testdata: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/crpdata_lv on /crpdata: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/testarch_lv on /testarch: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/locdata_lv on /locdata: There is a request to a device or address that does not exist.
mount: 0506-324 Cannot mount /dev/visdata_lv on /vision: There is a request to a device or address that does not exist.
Multi-user initialization completed
0

-t
maxclient
838008

number of valid memory pages = 1048537 maxperm=79.9% of real memory
maximum pinable=80.0% of real memory minperm=10.0% of real memory
number of file memory pages = 5290 numperm=0.5% of real memory

number of compressed memory pages = 0 compressed=0.0% of real memory
number of client memory pages = 0 numclient=0.0% of real memory
# of remote pgs sched-pageout = 0 maxclient=79.9% of real memory

vmtune: new values:
-p -P -r -R -f -F -N -W
minperm maxperm minpgahead maxpgahead minfree maxfree pd_npages maxrandwrt
104853 838008 2 32 120 512 524288 0

-M -w -k -c -b -B -u -l -d
maxpin npswarn npskill numclust numfsbufs hd_pbuf_cnt lvm_bufcnt lrubucket defps
838830 32768 8192 1 200 80 9 131072 1

-s -n -S -L -g -h
sync_release_ilock nokilluid v_pinshm lgpg_regions lgpg_size strict_maxperm
1 0 0 0 0 0

-t*******************************************************

i have checked for errpt log for errors and i got error messages array configuration changed.i have discussed the problem with one of our senior team leaders and he told me that the reason for the problem is difference between the kernels of aix 4.3.3 and aix 5.1 and this problem can only be resolved if i upgrade the s1 server to aix 5.1.this is only due to compatibility issue between aix 4.3.3 and aix 5.1.pls let me know what my team leader said is true or false??is there any compatibilty issue between aix 5.1 and aix 4.3.3??any patches available to overcome this problem.?pls let me know if u need any further details.

regards

arvind shrivastava
 
Arvind,

The error message &quot;There is a request to a device or address that does not exist.&quot; is normally seen when you attempt to mount a filesystem before performing a varyon of the volume group. Given that in your case the VG is set to autovaryon this implies that there is something wrong with the underlying devices and the VG cannot be varied on.

I am not aware of any kernel differences between 4.3.3 & 5.1 which would cause the problems you have described. There may be some LVM issues which would be relevant but these would only be for VGs which are using specific aix5 functions.

This is looking increasingly like a problem with the device drivers. I do not understand why cfgmgr is not finding the devices correctly on reboot but is finding them when you run cfgmgr using the CD.

Questions.
Which device drivers for the external disks are you using on server s1, which device drivers on server s3?

Does cfgmgr find the devices correctly if you do not use the CD? If not can you use the -v flag (verbose) on cfgmgr to find out what is being loaded from the CD?

Do you have any further details on the &quot;array configuration changed&quot; message you are seeing in errlog? What is the text of the messages, does this appear on every reboot of the s1?
 
when i run only cfgmgr command on s1 and s3 servers it gives an error message devices.fcp.array not found.but we have devices.fcp.disk.array file on s1 and s3 server.even when i run cfgmgr -i /dev/cd0 i get the same error message.yes on every reboot i get the message array configuration changed.
 
Again, what is the text of the error message (use errpt -a). Sometimes these messages are just informational or warning messages (you used to get these on VSS which could be safely ignored).

If this is a device driver problem then your best course is to raise a call with IBM support. The base issue appears to be that these disks are not being acquired automatically upon reboot and there are 3 basic reasons for this:

1) The disks cannot be acquired because of hardware issues (invalid cabling etc).
I do not think this applies to your case.

2) The disks cannot be acquired because the system lacks the required drivers.

3) The disks cannot be acquired because they have a SCSI reserve lock set.


 
Hi,

The reeason might be device driver problem but i am unable to find devices.fcp.array fileset in aix cd and fibre channel adapter driver cd.how can i check SCSI reserve lock set???

arvind
 
arvibm,

I am not aware of any fileset called devices.fcp.array, only devices.fcp.disk.array fileset which is on the base CDs.

I think you are going to have to raise a call with IBM. Either the problem is that you cannot acquire the disks because of device driver problems or the scsi reserve lock is not being released by the fastt500 when the s1 server shuts down.

I do not know of a way to check if scsi reserve lock is set other than the brute force method of powercycling the disks (if the lock is set then powercycling the disks should clear it). It is not recoomended that you do this. Call IBM.
 
I have logged the call with ibm and i got reply that this is an compatibility issue between aix 5.1 and aix 4.3.3 and this problem can only be resolved only if i upgrade my s1 server to aix 5.1


arvind
 
arvind,

Are they now saying that a fastt500 cannot support AIX servers on different levels of AIX?
 
Hi,

No they have not made any comments on FASTt500 support on different levels of aix.according to them i can not aix 4.3.3 and aix 5.1 to share the fastt500 but even they are not sure about the reason.no documents are available on net on compatibility of aix4.3.3 and aix 5.1.if u need some more logs from s1 server to reach to exact cause of problem pls let me know.according to ibm support this problem can only be resolved only if i upgrade the s1 server to aix 5.1.

thanx and regards

arvind shrivastava
 
arvind,

Was the s1 server down when the upgrade of of s3 server was done?
Are there disks on the s3 server in a defined state? (lsdev -Cc disk).

All I can think of was that something was written to the disks during the upgrade or during the actions to install drivers on the s3 after the upgrade which is preventing s1 from automatically acquiring these disks.
 
output of lsdev -Cc disk on s3 server
# lsdev -Cc disk
hdisk0 Available 40-60-00-4,0 16 Bit LVD SCSI Disk Drive
hdisk1 Available 40-60-00-8,0 16 Bit LVD SCSI Disk Drive
hdisk6 Available 21-08-01 3552 (500) Disk Array Device
hdisk7 Available 21-08-01 3552 (500) Disk Array Device
output of lsdev -Cc disk on s1 server:-
# lsdev -Cc disk
hdisk0 Available 40-60-00-4,0 16 Bit LVD SCSI Disk Drive
hdisk1 Available 40-60-00-8,0 16 Bit LVD SCSI Disk Drive
hdisk2 Available 21-08-01 3552 Disk Array Device
hdisk3 Available 21-08-01 3552 Disk Array Device

at the time of upgradation of s3 server s1 server is up and running.i have rebooted the s1 server next morning .u might be correct because before upgradation on s3 server the output of lsdev -Cc disk was different i.e. in place of hdisk6 and hdisk7 we have hdisk2 and hdisk3 but after upgradation of s3 server i have to reinstall the fibre channel adapter driver to configure FASTt500 disk and even after that s3 server is picking up wrong disk i.e. hdisk5.i have deleted the hdisk5 and used redefinevg -d <vgname> command to bring the volume group back on s3 server.when i compared the vg identifier number on harddisk on two servers i.e. s1 and s3 servers the vg identifier of hdisk2 and hdisk6 are same and hdisk3 and hdisk7 identifier are same.pls let me know if u need some more details to get the resolution for this problem.i got to know from ibm that this problem is new and nothing is written anywhere about this compatibility issue.
 
arvind,

Sorry, but I'm stumped. Looks like you'll have to wait for a solution from IBM unless anyone else here has any more ideas.
 
Hi,

I don't think any of the tek-tips members have any ideas to resolve the problem as u r the only person who has picked up my question and atleast tried to reolve the problem.i really appreciate ur attitude.
i don't think i will get any answer from ibm for the problem as they have already told me that the problem is new and nothing is written anywhere about it and as per ibm advise the problem can only be resolved by upgrading s1 server to aix 5.1.
i am new to this forum,can u pls forward this question thread to experts of the forum if they are ready to accept the challenging problem.
if u need any documents or problems realted with error codes then pls let me know.u r such a nice person i have ever met on net .thanx for ur help and support.

thanx and regards

arvind shrivastava
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top