Data missed when lengthy archive read occurs on start-up

[freddie]

MX version 3024. Vantage Pro2 via Virtual Vp (over TCP/IP). 1 minute logging interval.

I had another instance of data flat-lining yesterday (see separate thread). Stopped MX, copied last good back-up (taken at 19/0900) to data directory. Restarted MX at 1004 UTC today. Archive data read from 0900 UTC on the 19th until 1006 UTC today, which took until 1010 UTC. Live data then read from 1011 UTC today - therefore I have missed 5 minutes of data. Is there any way of getting this data back, other than watching the logs and performing a re-start as soon as the long archive read has occurred?

[steve]

This has always been a 'feature' of Cumulus, because of the way that it has a 'reading archive data' phase, followed by a 'reading live data' phase which lasts for the rest of the session. Someone did bring this up recently, and asked if it could have another pass at reading the archive data if the first pass took a while, and that's something I could look at in the future.

[freddie]

This has always been a 'feature' of Cumulus, because of the way that it has a 'reading archive data' phase, followed by a 'reading live data' phase which lasts for the rest of the session.

Ah, I didn't realise that.

Someone did bring this up recently, and asked if it could have another pass at reading the archive data if the first pass took a while, and that's something I could look at in the future.

That would be really useful.
Thanks for the reply.

[steve]

Actually, it's quite trivial to unconditionally do a second pass (for Davis stations only) so I'll give that a try. If there happens to be no more data anyway, the second pass should only take half a second or so.