iscsi deploys are sensitive to timeskew

[Robert Collins]

With the PXE driver, if the deployed-to node and the conductor are too far apart clock wise, errors that may be quite inscrutable will be thrown.

We should:
a) in tripleo ensure ntp is available on the conductor
b) have the deploy ramdisk try (with a timeout) to sync time with the conductor
c) document this issue for other deployment approaches

[Dmitry Tantsur (divius)]

Hi! Just for future reference: could you give some example of errors that can be thrown?

[Steven Hardy (shardy)]

This bug was reported against an old version of TripleO, and may no longer be valid.

Since it was reported before the start of the liberty cycle (and our oldest stable
branch is stable/liberty), I'm marking this incomplete.

Please reopen this (change the status from incomplete) if the bug is still valid
on a current supported (stable/liberty, stable/mitaka or trunk) version of TripleO,
thanks!

[Pavlo Shchelokovskyy (pshchelo)]

While not strictly related to iscsi deploy and albeit probably being obvious, I'd like to add that the whole Ironic service is quite sensitive re time skew between ironic-conductor and ironic-api nodes.

This is due to ironic-conductor nodes posting DB keepalives in their local time (not the DB time), so api service might decide that given conductor is out if the time skew is big enough (with default settings, >50sec).

The resulting error manifests as "No valid conductor host found for driver..." etc when e.g. trying to create a node.

[Jay Faulkner (jason-oldos)]

Marking this as a duplicate of our longstanding IPA bug to make the system clock accurate. I'll add tripleo as effected by that bug.