This archive is retained to ensure existing URLs remain functional. It will not contain any emails sent to this mailing list after July 1, 2024. For all messages, including those sent before and after this date, please visit the new location of the archive at https://mailman.ripe.net/archives/list/ripe-atlas@ripe.net/
[atlas] Dead probes - slightly OT
- Previous message (by thread): [atlas] Dead probes - slightly OT
- Next message (by thread): [atlas] Dead probes - slightly OT
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Philip Homburg
philip.homburg at ripe.net
Thu Oct 29 09:24:46 CET 2015
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015/10/28 17:50 , Gert Doering wrote: > On Wed, Oct 28, 2015 at 05:30:49PM +0100, Wilfried Woeber wrote: >> This seems to imply that the probes can bootstrap (off the 'net?) >> without the USB stick present ? If this is the case, what is the >> USB stick used for? > > As far as I understand from the "firmware upgraded"-messages, the > internal flash has some sort of immutable emergency recovery > firmware on it, which is good enough to pull an upgrade from the > master, install to USB stick, and reboot into it. Or so. > > And, for measurement data :-) Close enough :-) The tp-link comes with 4 MB flash. This not enough to store the measurement code let alone also storing measurement results as well. Worse, if a power failure happens at the wrong moment during an upgrade of the internal flash then the tp-link is bricked. So the architecture we came up with is that the tp-link boots from internal flash and then either sets it root directory to the USB flash stick and runs from that or if an 'empty' USB flash is detected, downloads firmware from a controller and writes it to the flash. This way, updating the code on the USB flash is completely safe. If something goes wrong it is just a matter of asking the probe host to perform the procedure to get the USB stick to be re-initialized. We can and have updated the firmware on the built-in flash. But that's quite rare and there is the risk that it may brick some probe. Philip -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJWMdfOAAoJEPr6076EDopy8QUP/0WF3Jls3Ilz7QYRs4gCegbI VI6yUsOlrLUTWpMsjmqhFML/1c3s+dUdkRnHg5jL4RvTLsUgDdW38O5VSoPEah6P HF9jIyUS2rIHz0H3G2pk3rCkjso+r9FI1W+BLYCvb0Fk2qBlkTFjVYIPHqFQITbz gJ+pz+H1PHpzTBpZtCcGOQvIeSXE3BV4sU8dTKSU+EcCOl5eSCSu9cAP+OgrHWiw iL+T7RLxIn2BcjAHKljgFh15vHaoH3FN7Egc1DoixozmVw242LpiCUqdc/FvORlp X75I32KJr8jb/0GSQF8NTzNbRVd6tMIQCGLgRQYXjWMg3Ig0D5QUSrZ5yHpg6XuW wBJGE1obw9P8wXwLUW8ZHu3tqmmx2OCKFluQwUzhKgH7TmxIfuQcA2zyCYGYaAIO oLDl49umrR3q4GRVZhTeC6SVpx6tK4U2kBeZjYgu/XOmYNj76sfJgekcXjMa6Hsn /x8u5mIQSJhiINa1ovm8QaxCOR2zX1BiphqlP8NGT/dtki6CGyvuKoHg1yrYqcnt aYXzt4xZS9kiwV/jS/cF7Y+SgaAwAovCit5UyukWeohhJw1ZYDai0WRGLHd5hqw0 bZnIPjawDPDbnWCVMfcBBMu/SxsQzEuqkm8cWHkEwfkGRmViAcpw0x7HwjsTpfXA WuptOiuF3CHaFRKn+IKf =DUtO -----END PGP SIGNATURE-----
- Previous message (by thread): [atlas] Dead probes - slightly OT
- Next message (by thread): [atlas] Dead probes - slightly OT
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]