Aggregate data not being built following rebuild of NLSU **

[iasgoscouk]

This morning i installed the 2.6b experimental update, and for some reason upon restart, the /data/ directory was empty.

I have performed a complete rebuild og 2.6, and copied the /data/weather /data/graphs back, and rebuilt the graphs upload.

However, I've noticed that the raw data is appending to the existing file, but the agregate stats under /data/weather is not being updated and therefore the graphs have missing data since.

Previous to the update I took a copy of all files under /data at 10:31 this morning.

I asked for a recompute about an hour ago, but nothing is happening.

[skyewright]

I asked for a recompute about an hour ago, but nothing is happening.

On the System Information tab does it say that recomputation is "in progress".

I install (cleanly so far as I can tell) and set Recompute going 4 hours ago. It's still "in progress".

That's the longest I've ever seen.

The system load is varying, and doesn't seem anything dramatic.

I have WD shutdown during recomputer (so that the logging protocol isn't increasing the load).

I just logged on to nudge the wish for some progress feedback on Recompute.

PS. Just under 5 months of data, with 'Storage' reporting:
Data: 501MB of 3180MB used (15%)

Later PPS. N.B. Not complaining or ranting about Recompute, just using my case to illustrate that Recompute can take quite a while (and did eventually finish - see below...).

[iasgoscouk]

the status says 'completed'; only have one months data.

tried everthing so far but nothing will start the creation of /data/weather data.

[skyewright]

...set Recompute going 4 hours ago. It's still "in progress".

Completed in around 4 and a half hours. Always a relief to see that "completed" appear. :)

asgoscouk wrote

the status says 'completed'; only have one months data.

Did you see it say "in progress" at all?

[iasgoscouk]

suddenly sprung into life about 15 mins ago, and most of the files have been updated

[admin]

When you press recompute, meteohub sets a flag that a recompute should be done the next time meteohub is looking for recomputation tasks. this happens about every 5 minutes. So when you press recompute and have a look at the status page, it might say completed also it is not completed. It is just waiting for start.

I will change the message on the start page to something like "waiting" when have a recompute job scheduled.

[web_smith]

I seem to have the same problem.
I asked for recomputation of the aggregated weather data and nothing happens. The funny thing is, that I changed a stale configuration file - I removed a reference to a non-existent sensor - and suddenly the *-min1 files got updated. After that the files are not touched and further updated. All other aggregated data are updated with the weather data since my upgrade to 2.6.

The history:
My WMR928 wrecked end of June and I had to send it in for warranty exchange. Last Saturday I installed the weather station again and decided to do a meteohub upgrade from 2.3e to 2.6 using backup/restore and new USB stick images and directly to 2.6a. All historic data are restored, but the aggregation fails - or does not start. No trace in log files apart from a non-existend sensor for wd - which I do not use.

Please advice for the aggregation of all historic data.

Stefan

[iasgoscouk]

I haven't received a technical reply as to why/what caused my aggregate data to stop building. I've changed the crontabs for my graphs to only run at certain hours giving a 'window' of 6 hours during the night for anything to catch up.

Following a rebuild of my NLSU and restore of data it took some 3 hours before the aggregate data started again, on its own.

I've requested a recompute;

The aggregate data is now adding to the files since then, but have a gap of 4 hours in the files which have not been recreated.

I had given up on that data, but i think a more technical explination of this should be given as to what causes the aggregate data not to be restarted following a restore.

All my aggregate data is now building except those beginning 'seq'.

IAN

[admin]

stefan, please press "save" on the sensors page, this will update the config files that control build of aggregated data.

[iasgoscouk]

that sounds a little familiar - when the data wasn't being built I went through every screen and am sure I chose 'Save and restart logging' at some point that afternoon.

[admin]

When you update to 2.6d, you will get some more info ont the "system information" page. When it says "waiting", a recomputation job has be scheduled. When it says "running" you can see how much has been done and what still to do by a small progress bar.

Please also check your "meteohub log" if there are messages in there that some invalid data blocks recomputation.

[iasgoscouk]

13:35 - I have installed 2.6d (it still says 2.6c at the top of page).

The status of the recomputation says waiting. This must be from when I requested this a few days ago.

There is no reports of bad data blocks - the only problem for me is that there is a gap of 4 hours in the logs for 28th Aug from 09:30 to 14:30.

13:44 - Just rechecked system info and it now says 'in progress' - so will not post this until it completes.

13:56 - Reached 'all solid dots' on the progress bar and completed 13:58.

Checked the 'th0-min30' file and there is now only a gap between 10:00 and 12:30 which accounts for the 2 hours physical downtime of the NLSU, but has recovered a missing 2 hours.

many thanks