XenCenter will check host.patches_requiring_reboot when a host restart is due while a batch update is in progress. If the restart is not required, XenCenter will skip relevant PlanActions.
The intention here is to check this whenever a restart is due according to the original schedule - this way a completely livepatchable update can prevent the restart while the key batch hotfixing logic is not changed.
Signed-off-by: Gabor Apati-Nagy <gabor.apati-nagy@citrix.com>
If we are expecting to live patch, but at the end
hosts.patches_requring_reboot contains the patch, then warn the user that
live patching has failed for the host - and it needs to be rebooted.
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
Intermediate checkin. Currently has fake functionality, will always claim every involved host requires a reboot. If one host, uses ThreeButtonDialog to offer to reboot it, else uses a new RebootHostsDialog to let the user choose which, if any, to reboot. Uses the progress bar + log already on the page, which is probably the wrong design but works for now.
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
This commit implements the check for LiveUpdate capability of a patch. If
the patch is completely live-updatable, XenCenter will not restart the
host (but will do further checks once it has finished installing the
update - separate ticket)
Signed-off-by: Gabor Apati-Nagy <gabor.apati-nagy@citrix.com>
Now using the EscapeAmpersands string extension method. Also resolved the custom fields display issue in the same way.
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
The ticket said the character after the & became a hotkey (like labels etc do with UseMnemonic enabled) but I couldn't reproduce that - & symbols were simply removed from the string. The fix is to replace them with &&, so the string.format prints a single & as it should.
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
the checks before the page is loaded. Corrected the conditions for enabling the
Previous and Next buttons on this page. This commit fixes broken RPU wizard test.
Signed-off-by: Konstantina Chremmou <konstantina.chremmou@citrix.com>
Comparing uuids instead of references that likely will be changed after the Cache has been repopulated.
Signed-off-by: Gabor Apati-Nagy <gabor.apati-nagy@citrix.com>
- when the cd drive or the tools iso cannot be found, the action should fail.
- if the tools SR is broken, do not show the error on a pop-up because the UX is
bad when installation is attempted on multiple VMS.
- added one more entry to the list of possible tools iso names.
- small performance improvement.
Signed-off-by: Konstantina Chremmou <konstantina.chremmou@citrix.com>
Generalising to also support pools. New CheckHostPatchesRequiringReboot method with the same behaviour as old method (return list of KeyValue warning messages). Pool updates on Ely use the union of messages from this method instead of the old one. Also localised the "{hotfix} on {server}" string per todo (hence resx changes).
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
For servers running Ely or greater we now produce messages for patching requiring reboot using the host.patches_requiring_reboot field. The Pool_patch objects resolved are sufficient to produce the same messages. Note that some patches produce a message in the old API but not the new, this is intentional. Still to do: update pool updates in the same way,
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
Now behaves as expected - alert interval (top control) has lowest tab stop, all other groups incremented by 1.
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
Instead of adding a second table layout panel for the rubric label, I have extended the original table to include two new rows at the top - one for the label rubric, and one for the repeat interval controls (previously not in the table).
Signed-off-by: Callum McIntyre <callumiandavid.mcintyre@citrix.com>
