I don't care....
...What the sales rep told you! You should *never* run a production rootfs from a micro SD card!
Switch to Linux, they said. What could possibly go wrong? Quite a lot, if today's edition of The Register's Bork diary is anything to go by. Reader "Clive The Commuter" spotted a BOGOF (bork one, get one free) as a pair of digital signs at London's Marylebone station had a wobble. "Aw, Snap!" says a cheery edition of what we …
My oldest devices in service are 8 and 10 years old and still on their original SD cards. The newer devices are doing things that require more speed so OS on SD card would cause a noticeable slow down. If the task is not demanding, you buy quality cards and over spec the capacity OS on SD can be fine. If you buy cards from the supermarket the cards might last a few hours but I would not trust them that long with valuable data.
Make the on-card filesystems read-only, put all writable stuff into a ramfs, and there's no issue with using SD cards. Flash is destroyed by writing, not reading, and for something like this that’s just a thin-client to a web-hosted service, there’s no need for writable persistent storage on the device.
I would, however, have also designed this with a recovery image so that any corruption of that main flash could be blasted away easily by a quick dd, but as it’s a mobile unit anyway, access for repair is probably not as big an issue as for fixed stuff.
Switch to Linux, they said. What could possibly go wrong?
Well, you should not expect a different result when the same (intellectually chalenged) designers and maintainers are employed, regardless of underlying system. Doing fail-resistant systems is hard, very hard. Making systems fail-safe (or fail safely systems) is even harder, very much harder. Just look at rocket science... We still have some devices floating out in space, very far away at 150+ AU, and still sending data. So, it can be done, building a stable system.
I do not think that anybody is prepared to pay for that any more.
"We still have some devices floating out in space, very far away at 150+ AU, and still sending data."
Apropos of SFA, this comment led me to Voyager site to confirm the distance gap between the 2 Voyagers. I said out loud, "oh, Voyager 2 is only 124 AU away" then smiled at the absurdity of that statement. I also noted that Voyager 1 is now 20.75 light hours away - I hope that someone's planning a party for the light-day milestone. :D
A set of standards must be written up delineating maximum and minimum lnflated volume at a given pressure,rate of inflation using 'standard' lung power, hand pump or motorised pump,
Impermeability at atmospheric pressure at sea level, penetrative resistance and colour spectrum availability. Development and delivery times must be determined and samples provided all subject to tender, the decision to be based on a ratio of price/quality assurance........
and there's the procurement of the string,......
To be fair, a request for a FSCK doesn't necessarily mean drive failed. It could have been a power or software problem which caused a shutdown without syncing everything. So that could possibly be down to an OS error, although it could also be one of two hardware failures or a non-OS software failure.
In this case, it's almost certainly a drive or interface failure. The sign appears to have an NCQ-capable SATA drive which is returning bus errors to the block device layer in response to both read and write commands, leading to I/O errors in the filesystem layer. The automatic fsck run during the startup process (probably triggered by the filesystem having failed to unmount cleanly due to drive errors last time the machine shut down) failed because of the I/O errors so the init script is asking for human intervention.
the same state of disrepair as the signalling and points on the Metropolitan Line (over which Chiltern runs), and what should have been a 30-min trip home stretched to three times the
usual scheduled amount of time
It was quite usual for the Chiltern line in my day and it looks like nothing much has changed from 30 years ago.
I used them for 14 months around 2013 and they were pretty good at getting me between Banbury and Birmingham. There were a few issues but most seemed to be caused by sad individuals foisting their terminal unhappiness onto us commuters. My only gripe was that I liked to use Snow Hill (it being quieter and also the terminus) and if anything went wrong the service went no further than Moor St. On a couple of occasions that meant a mad dash from Snow to Moor via the shopping centre.
And is there anything worse for a commuter than getting an SMS at 1530 telling them there is a problem with their train?
But I was only once seriously late. They had to lay buses on one evening and I was nearly an hour late arriving home. But most of the time they ran a very punctual and pleasant service.
I was damn lucky though. I was made redundant the same week that there was a landslide near Leamington Spa. That must have been a very unpleasant experience for my ex-fellow commuters.
This post has been deleted by its author