Re: [gentoo-user] speedup remote portage

[William Kenworthy]

On 27/2/20 3:51 pm, Robert Bridge wrote:
>> On 27 Feb 2020, at 00:08, William Kenworthy <billk@×××××××××.au> wrote:
>>
>> Hi,
>>
>>     due to space considerations on my laptop I have moved portage onto a
>> network share (moosfs, mfsmounted) - slower but works fine.  However,
>> being a laptop trying to update/install when out and about is both very
>> slow and network intensive through a vpn - again, it works but is even
>> slower to the point of not always being practical
>>
>> Is there a way to localise/speedup portage scanning parts of the
>> update/install process?
>>
> I used to use a portage tree shared over NFS. One thing that is worth considering, if you haven’t, is having the remote host manage tree syncing. This would remove the need for the laptop to be involved in the sync.
>
> Also, I generally would try to avoid updating the system while out and about, so that when I am updating it is a local network operation. Obviously this won’t work if you need a package when out and about, but if you are local to the server every night or couple of nights, it would significantly reduce the pain of updates.
>
> Cheers,
> RobbieAB.
>
All good ideas - currently the portage tree is common to ~20 odd gentoo
systems, mostly similar hardware but also some vms's.  This is what I
have pointed the latop at.  I used http-replicator for distfiles for
years but its recently been deprecated for tree-cleaning (I think) so
there is now an mfs mounted common distfiles (also outside the portage
directory).  Same with the common package files - they are extensive but
live outside the tree in different folders depending on hardware type. 
I already have an arm host dedicated to updating portage and building
arm32 packages for -K install on compatible systems - that all works
great and has been running well for quite awhile.  Originally I used a
git synced portage master with local git syncing for each host.  I have
converted it to a single git synced portage mfs mounted read only via
mfs on each host which while much slower, its less fussy and works well
except when the laptop is off-site.

I was hoping there was some way to cache portage operations and reuse
them as doing a simple emerge -p world will take 30 minutes or more and
pull a lot of network traffic - and if something fails it starts all
over again.  Once the scanning is done actual merges are fine with only
a small time increase.

BillK