Re: Is there a risk for filesystem errors?

[Kai Peter]

On 2020-01-31 12:57, Ottavio Caruso via wrote:
> On 31/01/2020 10:22, Kai Peter wrote:
> > Hi,
> >
> > in a former question there was confirmed that the monitor command 
> > 'quit' is similar to unplug the power cable. Now, for modern operating 
> > systems it is recommended to shutdown them (and not to switch off the 
> > power supply). Thus the filesystem(s) will be unmounted cleanly to 
> > avoid file system errors.
> >
> > The question is: _Could_ it happen that I get filesystem errors if I 
> > use 'quit' in the monitor? Especially for ext4 and ntfs filesystems?
> >
> > I assume it _could_ happen, but can someone confirm please or not?
>
> Really, it's just like pulling the plug while the system is running.
> Not sure about ntfs, but journalled ext4 will give you some level of
> protection. fsck will kick in at boot time, if the partition is marked
> as (fs_passno) 1 or 2 in fstab. Will fsck recover bad sectors?
> Probably, but why risk? Just use system_powerdown.
>
> If the booted image is a cdrom or a read-only hd, then "q" won't
> (probably) do any damage.

Thanks for reply, as well thanks Garcia.

So maybe a bit background. IMHO the usage the word "quit" in the monitor 
is a bit dangerous. I usually redirect the monitor to a unix/tcp socket. 
To disconnect from such a socket I use (usually) STRG+C. That's ok so 
far. But I think it is a common case to use 'quit' or 'exit' to close 
the _connection_. Example: If I connect via telnet to a to a smtp server 
I type 'quit' to close the connection, but I do NOT kill/quit the 
server. This is different with the qemu monitor. Even if the _function_ 
'quit' is a required one for qemu I would like to have a different 
wording. Perhaps 'kill' would be an alternative?

Just my thoughts.

Kai

--
Sent with eQmail-1.10.3 beta - a fork of djb's famous qmail