(FACT-3160) Fix size calculations for AIX blockdevice and partitions

[loopway]

• disks.rb: TOTAL PPs = USED PPs + FREE PPs. Therefor adding FREE PPs to TOTAL PPs produces incorrect disk sizes.

• partitions.rb: Using physical_partitions (PPs) to calculate LV sizes are not suitable if there is more than one mirror (can be two or three). LPs * nr_of_mirrors = PPs.

• info_extractor.rb: Adjusted regexp because using property LPs also matched MAX LPs.

[puppetlabs-jenkins]

Can one of the admins verify this patch?

[CLAassistant]

All committers have signed the CLA.

[smortex]

I guess you'll have to update the tests accordingly in spec/facter/facts/aix/disks_spec.rb and spec/facter/facts/aix/partitions_spec.rb?

[loopway]

@smortex I don't think it's necessary, since the sizes in the specs are specified without using nr. of LP's and PP's. Tests would have only failed in previous versions if mirrored LV's have been used and/or there was free space (PPs) in the VGs.

[joshcooper]

@loopway thank you for your pull request. Could you rebase your changes on top of main?

[joshcooper]

retriggering actions

[joshcooper]

@loopway did facter 3.x (in puppet-agent 5.x) correctly report blockdevice_hdisk*_size and partitions./dev/*.size facts?

The reason I ask is because facter 3.x on AIX called odm to collect this information:

facter/lib/src/facts/aix/disk_resolver.cc

Lines 18 to 69 in 6d7ffc6

	disk_resolver::data disk_resolver::collect_data(collection& facts)
	{
	data result;
	vector<string> disk_types;
	auto pd_dv_query = odm_class<PdDv>::open("PdDv").query("class=disk");
	for (auto& pd_dv : pd_dv_query) {
	LOG_DEBUG("got a disk type: {1}", pd_dv.uniquetype);
	disk_types.push_back(pd_dv.uniquetype);
	}
	auto cu_dv = odm_class<CuDv>::open("CuDv");
	for (string& type : disk_types) {
	string query = (boost::format("PdDvLn=%1%") % type).str();
	auto cu_dv_query = cu_dv.query(query);
	for (auto& cu_dv : cu_dv_query) {
	LOG_DEBUG("got a disk: {1}", cu_dv.name);
	disk d;
	d.name = cu_dv.name;
	{
	string device = (boost::format("/dev/%1%") % d.name).str();
	auto descriptor = open(device.c_str(), O_RDONLY);
	if (descriptor < 0) {
	LOG_DEBUG("Could not open device %1% for reading: %2% (%3%). Disk facts will not be populated for this device", d.name, strerror(errno), errno);
	continue;
	}
	scoped_descriptor fd(descriptor);
	devinfo info;
	auto result = ioctl(fd, IOCINFO, &info);
	if (result < 0) {
	LOG_DEBUG("Ioctl IOCINFO failed for device %1%: %2% (%3%). Disk facts will not be populated for this device", d.name, strerror(errno), errno);
	continue;
	}
	switch (info.devtype) {
	case DD_DISK:
	case DD_SCDISK:
	if (info.flags & DF_LGDSK) {
	d.size = (uint32_t)info.un.scdk64.hi_numblks;
	d.size <<= 32;
	d.size |= (uint32_t)info.un.scdk64.lo_numblks;
	d.size *= info.un.scdk64.blksize;
	} else {
	d.size = (uint32_t)info.un.scdk.numblks;
	d.size *= info.un.scdk.blksize;
	}
	break;
	default:
	LOG_WARNING("Expected a Disk or SCSI disk device, got device code '{1}'. This is probably a Facter bug. Please report it, and include this error message.", info.devtype);
	break;
	}
	}
	result.disks.emplace_back(move(d));

But in my testing facter 3.x incorrectly reported the amount of used disk space, and therefore capacity. For example, given

# df -Im /    
Filesystem    MB blocks      Used      Free %Used Mounted on
/dev/hd4        2048.00    288.61   1759.39   15% /

Facter 3 incorrectly reports used as 2.00 GiB:

mountpoints => {
  / => {
    available => "1.77 GiB",
    available_bytes => 1898340352,
    capacity => "53.08%",
    device => "/dev/hd4",
    filesystem => "jfs2",
    options => [
      "rw",
      "log=/dev/hd8"
    ],
    size => "2.00 GiB",
    size_bytes => 2147483648,
    used => "2.00 GiB",
    used_bytes => 2147483648
  }

While Facter 4 correctly reports 288.59 MiB:

mountpoints => {
  / => {
    available => "1.72 GiB",
    available_bytes => 1844871168,
    capacity => "14.09%",
    device => "/dev/hd4",
    filesystem => "jfs2",
    options => [
      "rw",
      "log=/dev/hd8"
    ],
    size => "2.00 GiB",
    size_bytes => 2147483648,
    used => "288.59 MiB",
    used_bytes => 302612480
  },

[joshcooper]

@loopway I submitted a new PR #2633 with your original commit along with a test fix, so I'm going to close this PR. Thank you for your contribution!