Bug 701305

> Description of problem:
When start the virt-manager or click on refresh pool this error is showed in the log:

--snip--
May  2 09:12:11 ibm-x3755-1 libvirtd: 09:12:11.109: error : virStorageBackendMpathRefreshPool:321 : in virStorageBackendMpathRefreshPool
--/snip--


> Version-Release number of selected component (if applicable):
libvirt-0.8.1-27.el6.x86_64
kernel-2.6.32-71.el6.x86_64
device-mapper-multipath-0.4.9-31.el6.x86_64


> How reproducible:
Always


> Steps to Reproduce:
After the aliases configuration in /etc/multipath.conf, the Storage Pool was created with the following xml :

$ cat mpath.xml

<pool type="mpath">
<name>mpath-test</name>
<target>
  <path>/dev/mapper</path>
</target>
</pool>

And defined with :

# virsh pool-define mpath.xml
  
Everything seems ok and virt-manager now is able to active this pool and populated it with 'mpath devices'.
However, every time when I try to refresh this pool, the following error was logged by system: 

libvirtd: 16:19:22.686: error : virStorageBackendMpathRefreshPool:321 : in virStorageBackendMpathRefreshPool


> Actual results:
libvirtd is showing "virStorageBackendMpathRefreshPool error" when the pool is refreshed.

> Expected results:
libvirtd doesn't show "virStorageBackendMpathRefreshPool error" on logs.

> Additional info:

Source code: src/storage/storage_backend_mpath.c

According to the source code libvirt is getting problems when try to check the  virFileWaitForDevices(); line 321.

static int
virStorageBackendMpathRefreshPool(virConnectPtr conn ATTRIBUTE_UNUSED,
                                  virStoragePoolObjPtr pool)
{
    int retval = 0;

    VIR_DEBUG("in %s", __func__);

    pool->def->allocation = pool->def->capacity = pool->def->available = 0;

    virFileWaitForDevices();

    virStorageBackendGetMaps(pool);

    return retval;
}
-- /snip --

Here the function virFileWaitForDevices, source code: src/util/util.c
-- snip --
void virFileWaitForDevices(void)
{
# ifdef UDEVADM
    const char *const settleprog[] = { UDEVADM, "settle", NULL };
# else
    const char *const settleprog[] = { UDEVSETTLE, NULL };
# endif
    int exitstatus;

    if (access(settleprog[0], X_OK) != 0)
        return;

    /*
     * NOTE: we ignore errors here; this is just to make sure that any device
     * nodes that are being created finish before we try to scan them.
     * If this fails for any reason, we still have the backup of polling for
     * 5 seconds for device nodes.
     */
    if (virRun(settleprog, &exitstatus) < 0)
    {}
}