Slapgrid object makes a connection to master node immediately,
thus if you create a slapgrid object without checking pidfile and if
master node responds very slowly, you may get tons of stucked slapgrid
processes by cron and system may become unusable.

See merge request nexedi/slapos.core!560

Factor test partitions formatting into a separate method to facilitate
customising the formating step in tests that need to do so.

Now that `StandaloneSlapOS.format` calls `slapos node format`, remove
redundant operations in `StandaloneSlapOS.format`, such as updating
the slapproxy.

Always put IPv4 before IPv6 in partition address list, because that's
the existing order and some things rely on that order, such as tests.

A nice example of Hyrum's law in effect.

See merge request !566

   Destination Reference wasn't used
   Display "Request By", this makes easy to identity if this is a loose instance

   This makes navigation easier.

   It helps to know who requested the instance

See merge request nexedi/slapos.core!558

    Just invalidate the certificate login

    This drops certificate revocation, it only requires to invalidate the Certificate Login to make login invalid.

   revoke occur directly on the Certificate Login

   Other users wont have permission to cancel valid Certificate Logins, only Compute Node owners.

  Now we dont rely on the document reference anymore.
  Invoke generateCertificate on tests else the user isnt valid.
  Note:  Login name returns the Certificate Login reference, to keep
    backward compatibility with the API (and current usage) it user
    the actual user reference, so it is easier to identify which document
    set the access status.

   This issues SSL Certificates and use Certificate Login, rather them require to revoke it.

See merge request !503

   By using getModificationDate, it ensures that the data is set according to the transaction, since catalog can be updated since the transaction started leading to a recent etag using outdated cache from the transaction.

See merge request nexedi/slapos.core!559

   Display only the Software Releases and if they can allocate or not, since display individual computers are not exactly good.
   Filter compute nodes by allocation scope (not closed forever)

   Include few fields and adjust the listbox columns to be user friendly (Display title rather them id).

  No need set as manager just use average init script.

	linux-vdso.so.1 (0x00007ffca2156000)
	foo.so => not found

ssh-keysign is not readable by everyone so we cannot do "ldd" on it:

$ ls -al /opt/slapgrid/shared/openssh/f79ee5e4e4573183783541f2ad9133e8/libexec/
total 1332
dr-xr-x--- 2 slapsoft slapsoft     90 Jul 12 22:10 .
dr-xr-x--- 7 slapsoft slapsoft    131 Jul 12 22:10 ..
-r-xr-xr-x 1 slapsoft slapsoft 145560 Jul 12 22:10 sftp-server
-r-x--x--x 1 slapsoft slapsoft 506208 Jul 12 22:10 ssh-keysign
-r-xr-xr-x 1 slapsoft slapsoft 366736 Jul 12 22:10 ssh-pkcs11-helper
-r-xr-xr-x 1 slapsoft slapsoft 338064 Jul 12 22:10 ssh-sk-helper