Re: not yet perfect, but...

David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

David,
I love this physicist's approach, do we see what we think to see? Oh well.

Just to check, you /do/ have a tmp_directory entry in your recv.ini:

_____________________________________
[parameters]
tmp_directory=temp
_____________________________________

pointing to the same partition as all the other receive directories?
How could this box run for years without the tmp definition in the parameters
section - don't ask me why, when, how, this recv.ini file was ported from
installation to installation... At least, there was no tmp_directory entry.
Yes, it may be that you need to add this entry by hand.

Now, as in my special (really that special?) setup where I want metop and
MSG data to be on different disks (having the least possible data
movements and maximized storage space in mind), how do I achieve that? I
have two target partitions, but the tmp obviously has to be on only one of
them. If that is the case, then the final files will be composed on only one
partition but will be shifted to the other later on.

Alternatively, I could bring both together now on a single terabyte or larger
disk. What has more advantages?

This is the content of the present recv.ini file, and note that I added the tmp
pointer only minutes ago, being aware that it is only 21, half the truth... still
reflected in the missed packets rate.
[]
[EUMETSAT Data Channel 2]

target_directory=F:\received

temp_directory=F:\receiving\tmp
What I thought about more than once is to draft a graphical description of
the data path - describing the various instances where the data pass
through and what happens where... Kind of a workflow map.

Again and again, I am desoriented by the various receiving directories...

Cheers, and thanks in advance for a new re-railing...

U.
Ulrich, according to the 2.2.4a TelliCast Help information (viewed via the HTML shell), you need

tmp_directory=<directory>

So, for example:

_____________________________________
[EUMETSAT Data Channel 2]
target_directory=F:&#92;received
tmp_directory=F:&#92;receiving&#92;tmp
_____________________________________


You are correct that with the data on different disks, one common entry in recv.ini would not provide the best solution. By the way: I hope that is two different physical disks, otherwise you will have the disk head madly dashing between the two partitions on the one physical disk! Far better to have all the received data going to one partition, and the processed images on a /different/ physical disk. Can you not hear the noise the disk is making? <G> The hard disk of my receive-only PC is just 80GB.

Talking of recv.ini, it happens that mine has:

_____________________________________
[parameters]
file_delivery_speed=20000000
_____________________________________


So, broadly, I would have one large and one small HD, ~100GB for just the &#92;received&#92; directory tree, and 1-2TB for the processed and archived files. Topic for discussion - which HD should have the OS? I might suggest the small HD, but it would likely depend what else was on the system.

Thanks to the others for their input as well - Arne/Ernst's log file analysis program has help uncover anomalies in other areas. I'm sure that with Linux you also need to work out the data flow carefully for best results.

Cheers,
David
--
SatSignal software - quality software written to your requirements
Web: http://www.satsignal.eu
Email: david-taylor@...

Join MSG-1@groups.io to automatically receive all group messages.