I have googled around quite a bit and tried many different commands to get system information about my AIX 5.3 box but none of the commands I've used have given me quite what I'm looking for. I am interested in finding out the model of the motherboard, and amount of available RAM, the RAID setup (i.e. 1, 0, 10, 5, etc) and the partition setup (logica volumes, I suppose?)
I have used the lspv -l <diskname> command to get logical volume setups of the disks but I don't know if I'm doing this correctly. I am very new to AIX and don't know any of the basic commands to get information about the system. Is there a reference manual somewhere or can someone here point me in the right direction to get the info I am looking for? I'm no stranger to Linux but AIX seems to be a completely different animal.
This is quite hard, because most AIX systems nowadays are LPARs (DLPARs or even WPARs) and thus the systems are virtual in essence. In Linux terms a (D)LPAR is virtual system running on a hypervisor built in hardware, a WPAR is a paravirtualized system (think of something like a OpenVZ-VE) running in a LPAR.
Traditional R/6000 systems (the predecessors of todays p-Series systems, which would be something like 8 years old or older) had a "sysplanar" and an "ioplanar" device. These both would (more or less, I'm trying to accomodate to your terminology here) do what a "motherboard" does in a PC. You could test their attributes by:
lsattr -El <device>
Where <device> is one of the devices shown by lscfg , i.e. "sysplanar0". Also try listing the "vital product data" with:
lscfg -vp
Again: Do you want to get the RAM installed in the physical system or inside an LPAR? The former information you get on the HMC, the latter is a value to be explained: an LPAR can have a dynamically varying amount of memory: there is a "minimum", "maximum" and "desired" value. Without the minimum available the system won't start. If possible the the LPAR will try to get memory on top of the minimum until the "desired" amount is reached. More than "maximum" will never be apportioned to the LPAR.
You can find all these values in the LPARs profile(s - there can be several) on the HMC. To get the momentary amount allocated issue
lsattr -El mem0
on the LPAR. Issue (only as root)
svmon -G
To get some longer-term memory statistics, but read the man-page of "svmon" to interpret the figures. The numbers are memory pages, which are 4k in AIX.
There is no general answer to this as it is depending on the RAID adapter used and its driver software.
The LVM is very deeply integrated into AIX (unlike the Linux LVM, which is quite new). In fact i do not even remember any AIX version without an LVM and i work with AIX for almost 20 years now. I'll try to explain the concepts in short, you may want to detail what you exactly want to know if this isn't clearing things.
Unlike Linux AIX boots directly from a VG called "rootvg", which is always there. All the "core" filesystems reside there ("/", "/usr", "/home", "/var", etc.). Most AIX systems are set up so that the OS is in the rootvg and every application (along with its data) has its own VG.
The basis of all this are physical disks. Actually "physical disks" is a strectch because this could mean LUNs from a SAN too - never mind, real disks are also covered. All these disks (or LUNs or RAID sets, ...) have a device entry named "/dev/hdisk<n>", where <n> is some number. You can list these disks by
lsdev -Cc disk
To create a volume group (VG) you need at least one such disk, which isn't shareable with another VG - VGs always make exclusive use of disks. Once the disks are allocated to a VG they become "physical volumes" (PVs). You can see the disks (PVs) belonging to a certain VG with:
lsvg -p <vgname>
You can also issue "lspv" to see all the PVs and the VG they belong to.
When disks are added to VGs they are (ahem - virtually) minced to small pieces called "physical partitions" (PPs). The size of these PPs varies (starting at 2MB up to 1GB) and is part of the VG definition, which you can see by:
lsvg <vgname>
VGs once had a limitation of max. 32 PVs in a single VG and max. 1019 PPs per PV. Therefore the PP size played a vital role in planning for a VG, but today certain VG types ("big VG", "Scalable VG") overcome these limitations. Still most systems will be set up so that they adhere to these limitations even if they aren't in place any more.
Now that we have the diskspace organized in VGs we start creating LVs. LVs are created by allocating PPs to "Logical Partitions" (LPs). A LP is the same size as a PP and is disk space abstracted from mirroring effects: if you want an unmirrored LV you select "1 LV copy" and every LP is matched by 1 PP. Mirrored LVs will have a 1:2- or 1:3-relationship and the data of every LP will be held in 2(3) copies in different PPs.
When creating LVs PPs will be selected automatically by default, but it is possible to have several levels of control over the process: "inter-policy" and "intra-policy" controls the distribution over several PVs, it is also possible to directly control the PP allocation to LPs with "map files", which can also be created from LVs to see the current layout. (see man page for "mklv" and "lslv" for details)
List all the LVs in a VG with
lsvg -l <vgname>
The LV finally can be used in several ways: as raw device (LVs also get entries in "/dev"), as swap space, as a filesystem or as a logdevice (1 per VG).
If something needs clarification or greater detail just ask.
