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] V3 Probe: USB Flash Drive Filesystem Corrupted
- Previous message (by thread): [atlas] V3 Probe: USB Flash Drive Filesystem Corrupted
- Next message (by thread): [atlas] V3 Probe: USB Flash Drive Filesystem Corrupted
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Michel Stam
mstam at ripe.net
Wed Jan 11 14:17:35 CET 2023
Hi Simon, Quick update. The corrupt filesystem is signalled by the Atlas backend depending on the behaviour of the probe; The probe tries to connect home by way of the SSH over port 443, but also sends several “SOS” messages via DNS to signal its state. The combination of SSH successful or not, the amount of SOS messages and their type (filesystem read-only, USB inserted, backend contacted) together yields the tag corrupted filesystem. The software that runs on the internal filesystem does very little other than initialising/updating the USB stick to the correct firmware version and signal its state to the Atlas backend. The measurements etc are executed from the USB filesystem. Btw, can you say something about the blink patterns on the probe. See also here: https://www.youtube.com/watch?v=JBmXbOqcty4 How to Troubleshoot RIPE Atlas Probe Version 3 youtube.com Hope this helps :) Cheers, Michel > On 11 Jan 2023, at 12:39, Michel Stam <mstam at ripe.net> wrote: > > Hi Simon, > > I would expect a more recent USB stick to complete a write / read operation more quickly than one of its older counterparts. Thats not a guarantee, I know. Funny enough, I actually used an old stick on one of my development V3 the other week and was surprised how slow the performance was (but the system did still work). > > When the V3 probes have issues, usually its either: > Power supply failing > Cabling failing > USB stick corrupt (unlikely after 3 units). > > Working setups sometimes degrade over time. Something I’ve seen with a USB analyser a good few years back is that some USB sticks don’t report write errors when the stick dies. > > I will dig a bit where the flash filesystem corrupted tag comes from and get back to you. > > Can you message me the probe ids that are causing issues? (Privately if you want to keep it off-list). > > And if you have the possibility, please humour me and try a completely different power supply and cable. It has resolved issues for me before. Just to get it out of the way. > > Regards, > > Michel > >> On 9 Jan 2023, at 20:36, ripe.net at toppas.net wrote: >> >> Hi Michel, >> >> Thanks for you answer. I'm using an Ikea KOPPLA (204.150.27) which provides 17W. It's powering two v3 probes and one v4 probe. The other probes work well, so i don't think that the power supply is causing this. >> >> What i meant was, that maybe some USB sticks are to slow and thus not able to perform writing of a big chunk of data in the expected time, and therefore the USB flash/filesystem is considered broken? Can you tell the exact cause for the "Flash drive filesystem corrupted"-tag is? What triggers this? >> >> I am aware that USB 3.0 is backwards compatible, but since USB 3.0 sticks are faster, maybe it makes sense to stick with them (hehe), to ensure that at least the full USB 2.0 potential is used. >> >> BR, >> Simon >> >> >> On 09.01.23 12:26, Michel Stam wrote: >>> Hi Simon, >>> >>> First thing that comes to mind to me, do you maybe have a power problem? If the unit receives just enough power to boot the base TP-Link, but not to power the USB stick this may trigger these problems. Would you be able to boot the unit with a different power supply (try 5W at the least or 10W), to see if this makes a difference? >>> >>> Speed shouldn’t be much of an issue, USB to my knowledge is backwards compatible. >>> >>> Cheers, >>> >>> Michel >>> >>>> On 31 Dec 2022, at 22:31, ripe.net at toppas.net <mailto:ripe.net at toppas.net> wrote: >>>> >>>> Hello, >>>> >>>> i have two v3 probes and i'm experiencing issues: "USB Flash Drive Filesystem Corrupted" >>>> >>>> I have tried multiple USB Sticks, including brand new ones and also sticks from "good" brands like Lexar and Sandisk. >>>> The issue is not permanent, the error appears and disappears from time to time: >>>> >>>> 2022-12-22 @ 12:14:04 UTC Probe auto-tagged Your probe #xxxxx was automatically tagged as "system: Flash drive filesystem corrupted" >>>> 2022-12-22 @ 18:14:42 UTC Probe auto-untagged Your probe #xxxxx was automatically untagged as "system: Flash drive filesystem corrupted" >>>> >>>> This leads me to the thought, that the flash storage is not the real problem. It's unlikely, since i tried multiple dongles. Could it be possible, that not the flash storage is broken, but the speed of the USB dongles is too low? >>>> >>>> All dongles I tried were USB 2.0, not 3.0, since the TP-Link TL-MR3020 is not USB 3.0 capable anyway. But maybe USB 3.0 is a better choice? >>>> >>>> What do you think? >>>> >>>> BR, >>>> Simon >>>> -- >>>> ripe-atlas mailing list >>>> ripe-atlas at ripe.net <mailto:ripe-atlas at ripe.net> >>>> https://mailman.ripe.net/ >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: </ripe/mail/archives/ripe-atlas/attachments/20230111/9e71d311/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: maxresdefault.jpg Type: image/jpeg Size: 104153 bytes Desc: not available URL: </ripe/mail/archives/ripe-atlas/attachments/20230111/9e71d311/attachment.jpg>
- Previous message (by thread): [atlas] V3 Probe: USB Flash Drive Filesystem Corrupted
- Next message (by thread): [atlas] V3 Probe: USB Flash Drive Filesystem Corrupted
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]