Sotadata wont load properly, doesnt seem to be showing any database information??
thanks!
Alan
Sotadata wont load properly, doesnt seem to be showing any database information??
thanks!
Alan
I also find the same. No data apparent.
73 Phil
seems to be missing here too… but I understand that there is some cunning code that (should) restart the processes within 15 mins… Fingers Crossed… Paul
Something going on. I can’t log any QSOs this morning.
K6YK
I think I broke it uploading my Lake District weekend logs! I’m not used to the new logging interface yet. I’ve decided to wait a while and try again later.
Not sure, been off a good half hour now for me… so that may not have worked…
Cheers
Alan
Hey guys,
Here I have a problem. When querying a SPOT, I click on the Summit to see the Locator. At that moment the system shows the Summit reference, but when I click it, it goes to my Sotadata profile and does not show the Summit data.
To include data I have no problems.
73
Carlos
PY2VM
It works again! Thanks to who did some magic in the background.
73 Heinz
YES!! Well done team… Were back online… hahaha
Thanks!
Alan
The “database” is actually a few software components, a web app, some middleware between that web app and the database engine and the engine itself.
We have seen the database engine try to allocate a huge amount of memory, fail and restart itself. When this has happened, it takes about 15mins for the other components to notice and fix their problems talking to the engine.
Andrew upgraded the server on Saturday so hearing there were problems was saddening at first. I looked at the database engine was working fine. I could use the console and talk to it. Also SOTAwatch was able to fetch summit info etc. What failed was the database web app and SOTAwatch’s ability to add chases to your logs. Something else was broken.
Whilst it’s nice to thanked, I deliberately did nothing. I just left it to recover itself. Which it did. It was in a “sub-optimal state” for about an hour.
I’m not sure if you need to refresh the page once it all restarts. I have seen missing data on my stats page after a restart that an nice thwack on F5 fixes. YMMV.
From what I can see there’s a couple of suspects, but it’s hard to determine what is symptoms and what is root cause. In this instance, the API went south, and we could start to see requests starting to take longer to complete. I suspect this ultimately causes the number of open database connections to starve until it restarts itself or people stop trying.
Of interest is that one of the culprits appears to be the RSS feed for spots, and that 94% of the inbound queries come from three IPs on the same subnet in Poland, with sub-30s update times. I do know who that is, because of course they log into SSO from that IP range too. Tempting to turn it off, but I can see there’s a couple of Wordpress plugins still using it (<2%).
The other culprit is the user count generation on SOTAMaps (that nice dropdown with flags and number of users). There’s a lot of caching there, so I suspect it’s when it finds a new IP and needs to do a geolocation look up that it slows down. Either way, whether it suffers from whatever else is going on, or is part of the cause, there’s a lot of variation in the run times but it’s consistently the longest running web request in the logs by an order of magnitude or more.
Work is underway to stabilise the DB API further, particularly by separating out the SMP and RSS stuff from normal SOTAData3 stuff.