According to the update from Atsuhsi it seems that IPMU have some constrains on ssh access.

So that LAM propose that the synchronisation comes from IPMU or PU to LAM and not from LAM to external site.
This access will only be using ssh and will be restricted to ideally only to the server that will pull the data.

We should agree on the frequency of the sync (ideally that could be done once a week during week-end)

What do you think?

Restriction from outside-IPMU to global network (by means which has non-RFC1918 addresses) in IPMU is somehow messy thing now and its coordination/operation could be changed in short time scale. Web access (80/443) is relatively less restricted than ssh (22), so one "reliable" possibility is webdav-rsync if we want to push from outside-IPMU. For getting from this archive, there would be bunch of ways from pick one file by one via web, etc.

Also if instrument development sites (like LAM, PU, ASIAA) are easily accessible from outside, ssh-rsync from IPMU archive server is one possibility. This will require some configuration at development sites, that ssh gateway is possible to reach local data storage at lab (or local archive?). How difficult depends on configurations at each site, so it is required to be checked site by site.

For frequency, I think once a week at weekend could work. For immediate data transfer from LAM to PU for short time analysis, we could do/plan transfer on demand as now we are doing.

yuki.moritani do you want to keep this open ? or shall we close it .

Since LAM data have been archiving, I close this ticket. Archiving JHU (or other institute) data will be handled under a new ticket, if we decide to do so.