RomanZ

Administrators
  • Posts

    3,588
  • Joined

  • Last visited

Posts posted by RomanZ

  1. @jmpr Sync has no any limitation for amount of files other than system resources. Every file / folder consumes some amount of memory (roughly - 1.5Kb) and some amount of CPU to be processed once requested, rescanned, etc.

    I suggest to start with simple checks: see if Sync process is still running. How many system resources it consume comparing to ones you have? If it does not run - try finding our if there are any core dumps in OS or which other process could kill it.

     

  2. @sklus You are running Sync manually not via package. It is totally okay, though will require a bit more efforts when upgrading: you'll need to upgrade your binary manually (well, pretty much just replace it). Also, running Sync from root isn't good idea too: by default, all files arriving to Sync will be created with root as owner and with root's umask.

    Ideally, run sync under your user or create a special one with next params:

    ./rslsync --config <your_config_file>

    Here is the guide to config. In config you likely want to specify storage folder, your port and NIC for webui and the listening port for Sync to transfer data.

     

  3. @zifnab Unfortunately, no big progress with Arch Linux for RPI1.

    Quote

     

    I don't mean to sound rude, but 4 months (reported in this thread in june, but reports started in may on arch & aur forums) that arch linux RPi1 users cannot use their backup server/solution (see the irony?).

     

    I understand your frustration about issue not addressed for 4+months. Though, it's not about time, it's more about priorities. The issue only happens on Arch Linux for RPI, and Sync works just fine on other distros (just tested with 2.24). We are small team, and we have to choose wisely where to invest our time. The Arch for RPI doesn't look to be very popular combo at the moment.

    Quote

    just for the record, the RPi 1 is a armhf board...

    RPI1 has ARMv6 architecture. Wiki says that "In Debian Linux, and derivatives such as Ubuntu, armhf (ARM hard float) refers to the ARMv7 architecture including the additional VFP3-D16 floating-point hardware extension (and Thumb-2) above." So for the sake of simplicity it is ARMEL, not ARMHF.
    I understand that likely RPI1 supports VFP-v2 but not v3 and technically it is possible to make a special build of pretty much any software that will utilize RPI1's floating point hardware. Although - see my note above about the time management for small teams.

     

  4. @BobAchgill Well, usually "instance" is a computer, but technically one can run 2 separate copies of Sync on a single computer, so I prefer the "instance" term.

    Ah, now I see your use case. You are concerned about the "owner" instance dying and inability to share new links. 2 solutions here:

    1. Use your external HDD (or just keep some backed up copy of your files somewhere). Don't use advanced folders, use standard ones. Keep the RW-key of the folder in safe place. If computer dies, you can setup a new instance easily using old RW key and data from external HDD - it'll keep seeding as previous computer did.
    2. Setup some Sync somewhere, keep it linked to My Devices of the owner, but keep all the folders in disconnected state. Keep data on external HDD or backup it to safe place. If computer dies, use the backup instance (it'll be the owner), just manually transfer the data to backup instance and connect the folder there.

    I initially thought that you are going to switch the HDD there and back on the fly, which is bad idea.
     

  5. @colink The article authors assume that anyone following have some basic Linux skills indeed.

    Quote

    eg "Installing Sync on Linux" starts with:" If you have btsync package installed",

    If you don't - ignore it. It's more for users making an upgrade from previous versions of packages.

    Quote

     

    Create file /etc/apt/sources.list.d/resilio-sync.list + more instructions.

    I assume many users will not already be using btsync so I presume I can ignore this part

     

    This one goes in "For Debian-based linux...", why should you ignore it if you are on Debian?

    Quote

    My linux host has a folder named etc/  is the instruction telling me to make folders named  apt/ and sources.list.d/ then create the file named /resilio-sync.list?

    You should have /etc/apt/sources.list.d/ path already, unless you are running rather old linux. If you don't - I'd advise to see man on sources.list.

    Quote

    wget -qO - https://linux-packages.resilio.com/resilio-sync/key.asc | sudo apt-key add -

    How?

    Just run it in console. If you've got remote machine - SSH to it and run command. Not sure that I understand the question here.

    Quote

    Will Risilio allow me to do an automatic or manual FTP sync from my desktop to my web host?

    Nope. Sync only works on it's own P2P protocol. It can't deliver or get files over FTP.

    Quote

    UPDATE: my web host (shared hosting) has confirmed that I will not be able to install Resilio on the server.

    That's rather strange. We've got several Syncs running on our EC2 instances with no issues. That would be a bit more complex if you'd like to store lots of data in S3, as S3 is complex to connect as a local filesystem, but if you got not that much of a data - setting up Sync on EC2 doesn't differ from any other remote linux machine.

     

  6. @BobAchgill Is your use case to store synced folder on removable media and use 2 different Sync instances to actually sync files there? If yes, 2 comments on the use case:

    1. Links are tied to Sync instance. So, if some link is produced by instance A, only A and it's peers that are in My Devices list will be able to serve it.
    2. In general, it is bad idea to push the same folder for 2 different Sync instances. Every instance keeps its own database corresponding to data it syncs. So every time you switch, new instance will see that some data in DB does not correspond to actual data.

  7. @tehpeh The "docker" way to update would be to kill existing 2.3.8 container and deploy a new Resilio Sync image. Use the same paths you used to run BitTorrent Sync image, so new container will pick up all the data and storage folder. I just tried it in my lab with no issues.

    Could you please share more details? I'm especially interested which docker images you used (there is an unofficial one for 2.3.8), which commands did you ran?

  8. @Manu.74 Okay, so you are using official package not on RPI1. upgrade is rather straightforward for you: follow this instruction. It is very similar to what you had before, but note that URL and package name has changed. Once you'll install new package, it will automatically migrate all your settings from the old one.

    @aosaigh Kind of. The package you refer to is an old "BitTorrent Sync". Now you need to install "Resilio Sync" package (see my response to @Manu.74 with the link to article with instructions). There is no update for existing BitTorrent Sync package (and not planned, as we are migrating to Resilio Sync).

  9. @petson (I suggest we are talking about documents corrupt, not the executables infected) Sync copies file to archive every time the file gets changed. So, the amount of copies in Archive depends on how many changing iterations were done on a file by you and by cryptolocker. I'd advise to sort and choose files by modification date: if you know for sure when cryptolocker started it activity, you can get non-encrypted files before.

  10. @jasondunn Archive is a good safety net for many cases: when user doesn't fully get the idea of synced folders, for companies (or just teams) that collaborate on same files, etc. Therefore we prefer to keep Archive turned on by default.

    In some cases (for example, if working with video files or another bulky content) it'll indeed occupy some extra space. Although, as @GreatMarko mentioned, it can be disabled for use cases like yours. And no, we don't delete Archived files intentionally: users might need them.