Thanks Thomas. I was worried the answer to my first question was going to be no.
My fallback on the LAN-based database was going to be to use the rsync extension to copy the database from the production weewx instance to the development instance. A third option would be to mount the database on the production Pi onto the development Pi. Then the LAN instance is only being read. Since LAN access is only to a development instance, it would be no big deal if it dropped some records in reports. But requiring a device driver nixes those ideas also. Does the driver need to generate ARCHIVE packets and try to store them? Or are just empty LOOP packets ok? I was actually planning to get LOOP packets from MQTT which my production weewx instance publishes. So, I wouldn't even want the development instance to grab LOOP packets other than the MQTT ones. phil On Monday, January 21, 2019 at 5:30:53 PM UTC-5, Thomas Keffer wrote: > > 1. Unfortunately, no. WeeWX expects a device driver. I suppose you could > create a dummy device driver that emits empty LOOP packets at regular > intervals, but I've never tried that. > 2. Would not make any difference. Sqlite does quite well locking around > reads and writes. > 3. The problem using a database on your LAN is that it exposes yourself to > network errors. WeeWX does not do very well recovering from these kinds of > errors. It won't crash, but it can drop a record. Still, lots of people > have used a database on another machine. Personally, I think the problem of > SD card writes is overstated. I have been running an instance of WeeWX > <http://www.threefools.org/weewx/status/index.html> on an RPi with a > Sandisk Extreme Plus SD card. It has been up for well over 4 years without > a disk crash. > > -tk > > On Mon, Jan 21, 2019 at 12:24 PM <[email protected] <javascript:>> wrote: > >> I am currently running weewx (on a Raspberry Pi) inputting data from a >> Vantage Vue station, using a sqlite3 database to archive the data. I >> publish the generated webpages via an external web hosting service. I want >> to set up a development weewx instance on a second Raspberry Pi and point >> it to the existing database to get generated output. This would allow me to >> have the output from my development instance display current data gathered >> from the production instance. But, if I hose my output while making changes >> in the development instance, I won't impact the published webpages. >> >> I'm looking for advice and have a few questions: >> >> 1. Is there a way to run a weewx instance that skips the device input >> (since that will be handled by the production instance) but goes through >> the output items a skin would do including generating html pages and >> copying them into the public_html directory? >> >> 2. Is migrating to MySQL a better choice for this kind of setup? I >> would think that using MySQL will solve any issues of locking caused >> if/when both instances access the database simultaneously (though one is is >> reading and writing while the other is only reading). >> >> 3. I am also thinking of moving the database (and maybe other weewx >> directories) to a Synology NAS and mapping those locations to my two weewx >> instances to reduce SDCard writes. Are there any significant gotchas or >> warnings I should take heed of regarding this? >> >> phil >> >> -- >> >> -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
