I put a 512 GB card in mine. Somewhat cheaper than a 1 TB card but still a good amount of storage for an offline copy of my multimedia content.
Storage is getting cheaper all the time. Probably by the time the OP gets per Librem 5, a 1 TB card will cost no more than what I paid for my card of half that size. So you can look on the bright side of the delays.
waiting for 3 years definitely doesn’t make me want to buy anything else from purism, I was considering buying a laptop, but this makes me look else where
I’m waiting for nearly 4 years now, since October 7, 2017. ANd I understand why it takes this time to design and build this type of device. When I now read in my German news that our authorities bought the Israeli spy software Pegasushttps://en.wikipedia.org/wiki/Pegasus_(spyware) I think it’s even worth to wait another 4 years. There will be nothing in the market which compares to the L5: OpenSource and with kill switches.
I guess that would come back to “it’s a regular Linux system”. What works works. Soft links. Hard links. Mount bind. Regular mount. Application configuration. It either works or it doesn’t.
You have to be careful with the timing i.e. if you are going to put something on an external drive then the external drive has to be available before the app is (in order words, the app has to be started after the external drive is) - or the app has to fail reasonably if the external drive is “not there”.
The good news is that it’s open source. So if some dubious app needs 100 MB of local storage and doesn’t work unless that local storage is on the root drive then anyone can do the work to make the app more storage friendly e.g. add new configuration option to move some of the local storage requirement onto a drive other than the root drive.
But isn’t that rather esoteric? Android adds extra layers to make such distinctions. But I’m pretty confident that there is no Linux app out there that weirdly examines it’s storage place. Given that on Linux an app and its data, config are traditionally separated and can well live on different mount points, that would really be weird.
However, even though possible, I think it will usually make more sense to move media and other data (which might include cached open street map data) to the SD card than apps.
Not only because apps are usually small, but also to ensure system stability:
My Galaxy S3 often lost the mounted SD, I assume due to vibration/shock. Doesn’t happen on my S5, but still… I’d never move binaries or essential file structures (e.g. /home/purism).
The system should stay fully functional without SD. @amarok
Yes, I didn’t mean that the app goes out of its way to ensure that its data is on the root file system.
I meant that, out of the box, the app puts its data in a directory that by default is on the root file system and the app provides no means to control that directly and it is not possible or not practical to relocate the entire directory (tree) to another drive.
At the very least, the fact that it is open source means that you can study the source to understand what directories are used and how they are used before actually moving anything.
I have had a chequered history of making directories remote without telling the application (i.e. making the directory remote behind the application’s back and hoping that the application doesn’t notice) NB: on desktop Ubuntu, not on Librem 5
e.g. application using remote directory worked with one version of Ubuntu but simply didn’t work after a Ubuntu version upgrade and
e.g. application does not gracefully handle the situation where the application is started but the remote directory is “not there” (such as that it failed to mount) - this is a classic situation for a failure mode that was not anticipated by the developer and indeed is “impossible”
I understand that failure modes for “remote” are more numerous than failure modes for “moved to another drive” - but the concept is similar and in fact Linux tries to make all three situations indistinguishable.
As noted above, on the Librem 5 right now moving /home/purism to the uSD drive does not completely work.
For sure, it’s an extra point of failure. Not just loss of mount but actual corruption. Still, Raspberry Pi computers manage to run off uSD where that is the sole drive.