sd getting hammered

[mvallevand]

I think the issue is a new 100 login restriction in 24 hours implemented Jan 7th. Sub, Robert said you were sent an email.

Here you can see the user was requesting new tokens because of failures.

2025-01-14 02:44:38.299 [INFO][51] SD got new token: 95daac8140e3bec615b1afb55f51b9d1
2025-01-14 02:44:38.881 [INFO][51] SD got new token: a22e0b7af2fb40690a6edd8777c4063f
2025-01-14 02:44:39.869 [INFO][51] SD got new token: 5476ecb44379735d8ed0b529e08db93d
2025-01-14 02:44:41.062 [INFO][51] SD got new token: c858389b2915c6d37930619d8a4dc36b
...

Which match the SD records were you can see the other server instance calls interspersed after the collisions.

| 95daac8140e3bec615b1afb55f51b9d1 | | 2025-01-14 07:44:39 |
| 1d34d0b9191c70a4493c4dd777d7b643 | 2025-01-14 07:44:39 |
| a22e0b7af2fb40690a6edd8777c4063f | | 2025-01-14 07:44:39 |
| 7b8721e8d9b51d1ff2f30fecd81d608e | | 2025-01-14 07:44:40 |
| 5476ecb44379735d8ed0b529e08db93d | | 2025-01-14 07:44:40 |
| 2d970491066aad932c442723e8583208 | | 2025-01-14 07:44:41 |
| c858389b2915c6d37930619d8a4dc36b | | 2025-01-14 07:44:41 |
...

Moving the EPG update time is the only solution.

Martin

[techpro2004]

oddly enough, sd is working again, don't know why. Just to be safe I set the update time 12 hr apart. Also some old lineups reappeared and I removed them. I do not use a vpn and my wan ip is pretty stable but still dynamic. thanks.

[mvallevand]

12 hours might be overkill if it moves into a time you are using NextPVR on either server.

SD confirmed your remote IP was fine, it was definitely the new 100 login issue,

Martin

[sub]

Sub, Robert said you were sent an email.

Yeah, I had replied to him, but he didn't come back to me.

[sub]

From what I can see NextPVR is generally good at requesting a token, and reusing it until it needs a new one. However, the issue I found was that when it is doing an EPG update, it does the SD channels in chunks of 10 channels at a time. If the token gets invalidated part way through that chunk, it asks for a new token and successfully receives one...but continues to use the old one for the rest of the chunk (with each call failing, and going through the process of asking for a new token each time). This can be a large number of calls for each, so potentially large number of requests for new tokens.

This logic nothing new to 7.x though, and has been the same for years. I suspect it's relatively uncommon for a user's token to become invalidated, but likely the change to their login limits has made it a bit worse, because it can't make it's way to the next chunk without being locked out for a period.

[mvallevand]

That fix will help a lot but once they start colliding after a 10 channel call and alternate servers like the SD logs show using a good or bad token won't matter.

Martin

[techpro2004]

Heard from sd overnight. they say my ip is not being blocked by their firewall. I still have my lineups listed in npvr so I am assuming it it still working. thanks.