Recently I'm taking my first steps in AIX, initially with no real issues. Until now, off course...
I hooked up a couple of iSCSI disks to the system, with a RHEL5 machine running tgt on the other side. When running cfgmgr after initial configuration, I got an error:
Code:
# cfgmgr -l iscsi0
cfgmgr: 0514-621 WARNING: The following device packages are required for
device support but are not currently installed.
devices.iscsi.array
I googled my ass off, but couldn't find anything, other than some chinese site, wich I find very difficult to understand since chinese is one of very few major languages they don't teach in dutch schools...
Nonetheless, I was able to see the devices that I offered through tgt, and I was able to create PV's on them:
Code:
# lspv
hdisk0 00c9a1fbe1fd35d0 rootvg active
hdisk1 00c9a1fbe5bf8aaf rootvg active
hdisk2 00c9a1fb88919112 None
hdisk3 00c9a1fb6f64e590 None
hdisk4 00c9a1fb6f652404 None
However, I'm not able to make VG's out of any of these PV's:
Code:
# mkvg -y foovg hdisk2
0516-008 /usr/sbin/mkvg: LVM system call returned an unknown
error code (-267).
0516-1184 /usr/sbin/mkvg: IO failure on hdisk2.
0516-862 /usr/sbin/mkvg: Unable to create volume group.
Though I am able to read off the disk with dd:
Code:
# dd if=/dev/rhdisk2 of=/dev/null bs=1024 count=1000
1000+0 records in.
1000+0 records out.
Also lsdev doesn't seem to have a problem with the disk:
Code:
# lsdev -l hdisk2
hdisk2 Available Other iSCSI Disk Drive
With errpt I see the following error:
Code:
#errpt -a | head -46 #-46 was exactly the lenth of the error,
#I made sure it covered the _whole_ error
LABEL: SC_DISK_ERR2
IDENTIFIER: B6267342
Date/Time: Tue May 11 13:15:34 CDT 2010
Sequence Number: 29
Machine Id: 00C9A1FB4C00
Node Id: watson
Class: H
Type: PERM
WPAR: Global
Resource Name: hdisk2
Resource Class: disk
Resource Type: osdisk
Location:
VPD:
Manufacturer................Areca
Machine Type and Model......AIX/System
ROS Level and ID............52303031
Serial Number...............
Device Specific.(Z0)........000005423D000002
Device Specific.(Z1)........
Description
DISK OPERATION ERROR
Probable Causes
DASD DEVICE
Failure Causes
DISK DRIVE
DISK DRIVE ELECTRONICS
Recommended Actions
PERFORM PROBLEM DETERMINATION PROCEDURES
Detail Data
PATH ID
0
SENSE DATA
0A00 2E00 0000 0008 0000 0100 0000 0000 0000 0000 0000 0000 0102 0000 7000 0500
0000 000A 0000 0000 2000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 1C00 0000 0000 0000 0000 0000 0000 0000 0083 0000
0000 0029 0017
I'm afraid that this is where I find myself to n00b to handle this, maybe you have some pointers for me to fix this?
This one on an example using CISCO hardware and a DS4000. As far as I could understand from other sites, the step where you have to define iSCSI targets in /etc/iscsi/targets and the chdevs etc. will be the same for other models of hardware: Implement iSCSI on AIX
Thanks for your reply!
All of these articles, except indeed for the german one, I already found and read. Thanks a lot anyway though! :-)
To make things more complicated (or clearer to some), I was able to create a filesystem on the disk:
Code:
# mkfs -Vjfs2 /dev/hdisk2
mkfs: destroy /dev/hdisk2 (yes)?
File system created successfully.
97652620 kilobytes total disk space.
Device /dev/hdisk2:
Standard empty filesystem
Size: 195305240 512-byte (DEVBLKSIZE) blocks
Maybe a detailed output for lsattr enlightens my situation to some:
Code:
# lsattr -l hdisk2 -E
clr_q no Device CLEARS its Queue on error True
host_addr 172.16.100.15 Hostname or IP Address False
location Location Label True
lun_id 0x1000000000000 Logical Unit Number ID False
max_transfer 0x40000 Maximum TRANSFER Size True
port_num 0xcbc PORT Number False
pvid 00c9a1fb889191120000000000000000 Physical volume identifier False
q_err yes Use QERR bit True
q_type simple Queuing TYPE True
queue_depth 1 Queue DEPTH True
reassign_to 120 REASSIGN time out value True
rw_timeout 30 READ/WRITE time out value True
start_timeout 60 START unit time out value True
target_name iqn.2010-01.storage.example:foohost:areca.raidset3 Target NAME False
Where [foohost.example.storage] has IP [172.16.100.15] and is said RHEL5 host running tgtd. This host also serves iSCSI disks to a HP-UX 11.31 machine, where it works just fine.
Sorry for posting twice in a row, but it seems more appropriate than an edit to me :-)
As good citizenship mandates, hereby I'm back with what seems to be a valid explanation. After all, it seems that it is not an AIX issue, but an issue of AIX in conjunction with stgt.
Apparently mkvg in AIX, and in AIX only from what I found out so far, uses a SCSI "WRITE AND VERIFY" command, also known as opcode 0x2E. This opcode is a rather obscure one, that's rarely used, and also one that is not implemented in stgt. Since, from what I was told, mkvg is the only command that uses this function, everything else works nice and dandy.
As it looks now, there is no possible way of making this work with the software as it is. I've yet to find out whether ietd suffers from the same lack. For that I _will_ use the edit function (if no one replied in the mean time) ;-)
I would like to add for the record that this is setting an excellent example on how to finish a thread. Great follow-up, mr. aart. I hope to see such endings to (resolved) threads more often.
Last edited by bakunin; 05-17-2010 at 10:40 AM..
These 3 Users Gave Thanks to mr.aart For This Post:
AIX 6.1's iSCSI-initiator works with IETd target [SOLVED]
At the risk of looking spammy, here my final remark on this subject: I've successfully ran 'mkvg' on an iSCSI disk, exported from an RHEL5
machine running IET (http://iscsitarget.sourceforge.net/).
Code:
# chdev -l hdisk2 -a pv=clear
hdisk2 changed
# lspv
hdisk0 00c9a1fbe1fd35d0 rootvg
active
hdisk1 00c9a1fbe5bf8aaf rootvg
active
hdisk2 none None
hdisk3 00c9a1fb6f64e590 None
hdisk4 00c9a1fb6f652404 None
# chdev -l hdisk2 -a pv=yes
hdisk2 changed
# lspv
hdisk0 00c9a1fbe1fd35d0 rootvg
active
hdisk1 00c9a1fbe5bf8aaf rootvg
active
hdisk2 00c9a1fbad59e652 None
hdisk3 00c9a1fb6f64e590 None
hdisk4 00c9a1fb6f652404 None
# mkvg -y foovg hdisk2
foovg
# lspv
hdisk0 00c9a1fbe1fd35d0 rootvg
active
hdisk1 00c9a1fbe5bf8aaf rootvg
active
hdisk2 00c9a1fbad59e652 foovg
active
hdisk3 00c9a1fb6f64e590 None
hdisk4 00c9a1fb6f652404 None
Thanks to everyone that helped me fix this really _really_ annoying
issue!
--edit--
Useful to add that (or at least so I think) is that the odious 'devices.iscsi.array' warning by cfgmgr no longer appears since using IET.
Hi all,
Is there anyway i can control the naming of the disk device ?
I have added an iscsi disk on server1 using iscsiadm , devfsadm and it is now showing as
/dev/dsk/c2t0d0 in node1
However, i am worried that the same disk will appear as
/dev/dsk/cXtXd0 in node2
This iscsi disk(lun)... (0 Replies)
Previous forum threads have noted that AIX cannot use iSCSI disks because it fails when trying to add them to a volume group using mkvg. The failure is caused by AIX's use of the SCSI "WRITE AND VERIFY"(0x2e) command which some iSCSI targets do not support.
We experienced the same problem using... (1 Reply)
Hello,
Images won't work on UNIX when I try posting them on my website I'm working on. It doesn't show the image, and it's simply erroring.
Help! Thanks! (5 Replies)
Hi! I'm sure that somebody here installed freeBSD from a download of a virtual disc (.iso). But I made 5 downloads of 5 differents freeBSD installation (and no one has worked).Can somebody tell me where to download and how (if needed) to prepare the cd? (8 Replies)
Hi,
I have a very frustrating issue! I hope you guys can assist
When a disk is presented out the iSCSI target display a lower disk capacity
SOLARIS VERSION is SOLARIS 10 05/09 Kernel Patch 139555-31
ISCSI Patch 119090-31, 141878-11
Unix Commands To discover Target
bash-3.00# i... (0 Replies)
I just installed a fork of opensolaris, and I really like it.
I would hate going back to Ubuntu, But one issue, I cannot get my Dell Wireless 1450 Wireless USB Adapter working.
On Ubuntu 10.4 I just installed nonfree firmware, but I guess it doesn't work that way on Solaris, any help would be... (5 Replies)
I typed:
echo "echo hi > at_log.txt" | at now +1minute
to test the at command on my terminal. I got the message:
job 8 at Thu Feb 25 18:42:00 2010
I waited for a minute but nothing happened. I tried listing at_log.txt, but it said there was no such file. Am I doing something... (2 Replies)
Hi All,
can anybody tell me what's wrong with this code?
# SEARCH replaced by REPLACE
#!/bin/bash
SEARCH="95$$ 0 t"
REPLACE="95$$ 1 t"
for I in `ls *000.inp | cut -c-12`;
do
echo $I
sed 's/$SEARCH/$REPLACE/' ${I}-000.inp > ${I}-100.inp
done
It don't replace the string... (5 Replies)
I'm trying to figure out why this if statement won't work:
if || $zipcount != 6 ]]
then
echo ${myline}
echo "ZIPCODE WARNING! ${zipcode} ${zipcount}"
fi
if ]]
then
echo ${myline}
echo "STATE WARNING!... (3 Replies)
hey programmers!
1-why won't gcc accept as an argument? i tried the recommendations on the man page of getch(),..etc. nothing worked.
2-why it won't see <iostream> && <fstream> even if i implemented the function as follow
std::cout<<"..etc"<<endl;
3-after i type this code in it gives... (6 Replies)