Setting up iSCSI drive on AIX 5.3 1

[ngxGraz]

Hey guys, I'm running AIX 5.3, and using a NetApp, where I set up an aggregate thats about 133 gigs, then made a 100gig volume in that (effectively 80 gigs) then i made an LUN in this volume the same size (not space reserved) and finally I made a group.
I used the instructions at

http://now.netapp.com/Knowledgebase/printsolution.asp?id=kb2064

and after a while I got it all connected. I still had to have it mounted. by playing with smit vg, lv and fs, I kinda got it setup.
vg called iSCSI, and the fs mounted it to /iscsi. I played around with the size but i couldn't get it bigger than about 10 gigs (in half kilobyte chunks is how it's defined?) which is problem number one.
I get back to my box (remotely) the next day, and it's no longer accepting connections, ping only responds every OTHER time.. so I go locally and reboot, same ping problem, but I could at least telnet in now. I look at the filesystem, and well it doesn't mount. lsvg shows the iSCSI volume group, but when remaking a logical volume (enhanced journalled) it doesn't show up in the choices (using smit) so right now I'm kinda stuck.

Anyone done this before/able to give me a hand?

Thanks in advance.

 

[mrn]

Sounds like a routing problem.

Mike

"Whenever I dwell for any length of time on my own shortcomings, they gradually begin to seem mild, harmless, rather engaging little things, not at all like the staring defects in other people's characters."

 

[RMGBELGIUM]

okay, first of all about problem 1 : are you using snapshots ? If so , what is the snap reserve you are using , snapshot schedule , and what is the fractional reserve ( vol options <volume> )?

problem2 : Are you using jumbo frames ? If its a routing problem, you wouldn't get any response.Issues with mtu size, can cause severe packet loss.

rgds,

R.

 

[ngxGraz]

Here's my snapshot s chedule (it was default when i made the volume using the gui)

vol2	hourly.5	Jun 23 22:00	128 KB	1020 KB	normal
	vol2	hourly.4	Jun 24 02:00	128 KB	892 KB	normal
	vol2	nightly.1	Jun 24 06:00	128 KB	764 KB	normal
	vol2	hourly.3	Jun 24 14:00	128 KB	636 KB	normal
	vol2	hourly.2	Jun 24 18:00	128 KB	508 KB	normal
	vol2	hourly.1	Jun 24 22:00	128 KB	380 KB	normal
	vol2	hourly.0	Jun 25 02:00	128 KB	252 KB	normal
	vol2	nightly.0	Jun 25 06:00	124 KB	124 KB	normal

and smitty chgenet says I don't transmit jumbo frames.
Not sure about the snap reserve though :s

 

[ngxGraz]

also, wehbn trying to view the vg, or make a lv, it says i need to varyonvg

gaat:guest:/> varyonvg iSCSI
0516-008 varyonvg: LVM system call returned an unknown
        error code (3).

I never had had that error before when i first got it working.

 

[ngxGraz]

netapptest> vol options vol2
nosnap=off, nosnapdir=off, minra=off, no_atime_update=off, nvfail=off,
snapmirrored=off, create_ucode=off, convert_ucode=off,
maxdirsize=10470, fs_size_fixed=off, guarantee=volume, svo_enable=off,
svo_checksum=off, svo_allow_rman=off, svo_reject_errors=off,
fractional_reserve=100

 

[RMGBELGIUM]

when you look on your netapp box, can you see the AIX host log in ? ( iscsi show initiators ).If you cannot see your AIX initiator, then there's a connectivity problem, or the IQN number of your AIX host was changed.First of all make sure that iscsi is started on the netapp. The varyonvg failure, probably due to an VG quorum that isn't met.

 

[ngxGraz]

No, iscsi show initiators only shows a previous iscsi connection. iscsi IS running too, and iscsi show initiators DID show the connection before (when i said it was ten gigs)

gaat:guest:/> chdev -l iscsi0 -a initiator_name=blabla
iscsi0 changed
gaat:guest:/> chdev -l iscsi0 -a initiator_name=iqn.1992-08.com.ngx:gaat.c6a1b87b
iscsi0 changed

netapptest> Mon Jun 25 10:00:00 MDT [kern.uptime.filer:info]:  10:00am up 33 days, 23:27 0 NFS ops, 832 CIFS ops, 1100 HTTP ops, 0 DAFS ops, 0 FCP ops, 507709 iSCSI ops
Mon Jun 25 10:00:00 MDT [sfu.firmwareUpToDate:info]: Firmware is up-to-date on all disk shelves.
Mon Jun 25 10:29:53 MDT [iscsi.textkey.invalidValue:warning]: ISCSI: iswta, Invalid value for iSCSI text key, 'InitiatorName=blabla'
Mon Jun 25 10:30:20 MDT [iscsi.notice:notice]: ISCSI: iswta, New session from initiator iqn.1992-08.com.ngx:gaat.c6a1b87b at IP addr 172.19.1.55

So it looks like it's connecting at least.

 

[RMGBELGIUM]

you should be able to see the disk with lspv then ( or after running cfgmgr ).If the disk shows up there , there is no reason for the VG not to varyon.

regarding the 10 Gb size, I think you could be hitting the limit in PP's in your volumegroup.If you make the volumegroup a Big VG, do you still have the same problem?

 

[ngxGraz]

no it's really weird, it use to show, but doesn't anymore.

gaat:guest:/> lspv
hdisk0          00cb273fe9dda016                    ngxshrv1        active
hdisk1          00cb273fe9ddbc3f                    ngxshrv2        active
hdisk2          00cb273fe9ddda69                    ngxshrv2        active
hdisk3          00cb273f5060359f                    rootvg          active
gaat:guest:/> lsdev -Cc disk
hdisk0 Available 05-08-00-3,0 16 Bit LVD SCSI Disk Drive
hdisk1 Available 05-08-00-4,0 16 Bit LVD SCSI Disk Drive
hdisk2 Available 05-08-00-5,0 16 Bit LVD SCSI Disk Drive
hdisk3 Available 05-08-00-8,0 16 Bit LVD SCSI Disk Drive
hdisk4 Defined                Other iSCSI Disk Drive

hdisk4 being the correct disk, obviously.
I tried setting it as a big VG, but I can't edit it or remove/re-add it.

0516-306 getlvodm: Unable to find volume group A in the Device
        Configuration Database.
0516-882 reducevg: Unable to reduce volume group.

 

[ngxGraz]

So reconfigured the hdisk4, made a new VG, new LV and new FS, and now it's 80gb (I set it to the size netapp shows, even t hough it's set at 100, it only effectively has 80, is this correct?
here are my settings for each creation (order as it is the mkvg mklv etc smitty form)
vg (Big)
iSCSI2
1024
hdisk4
yes
yes
46
no

lv
iSCSI2v
iSCSI2
1
jfs2
middle
minimum
1
1
active
yes
yes
none
512
parallel
no
none
not striped
no

FS (Enhanced Journalled FS)
iSCSI2
512 bytes
167772160
blank
yes
rw
blank
no
4096
no
0
v1
no
no

 

[RMGBELGIUM]

Well, the 80 Gb , although the volume you created was 100 Gb , is due to the fact that netapp uses a 20 pct snap reserve by default, so if you do a "snap reserve <volumename>" on the netapp box , you will see 20 pct or 20 Gb .If you do not use snapshots, just issue "snap reserve <volumename> 0", then you will have the full 100 pct of your volume.

kind rgds,

R.

 

[ngxGraz]

netapptest> Tue Jun 26 07:07:28 MDT [telnet_0:info]: administrator logged in from host: ngx-w062.ngx.local
snap reserve vol2
Volume vol2: current snapshot reserve is 20% or 20971520 k-bytes.

You're great, thanks a lot