agent fails to reconnect if adapter configured by hostname and not IP about cppagent_dev HOT 4 CLOSED

On Oct 11, 2022, at 13:23, MRIIOT ***@***.***> wrote: Chris Misztur 11:04 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665072285189109> What is the default reconnect interval if adapter is down when agent 2.0 starts? will_s 11:06 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665072404365129> 10 seconds I think. Chris Misztur 11:08 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665072490338179> It was kinda weird. Adapter was stopped so hostname could not be resolved by agent. I started the adapter and agent did not reconnect. I had to restart agent. I’ll try to reproduce. will_s 11:14 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665072882455529> If you can turn on debug logging, I wonder if the name resolution is not completing and hanging to connection. Everything is asio, so the first step to connect is to resolve. Usually comes right back with success or failure. I may need to add another timeout. I would give it a little longer and see if the resolution times out. On airplane now. Will check when I get back Chris Misztur 11:26 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665073592216139> Here is at Agent startup with adapter off. 10/06/2022 11:25:01 AM 2022-10-06T16:24:59.095724Z (0x0000007fa2345fc0) [error] MAIN->Agent::start->Connector::resolve: Cannot resolve address: fanuc_driver:7878 10/06/2022 11:25:01 AM 2022-10-06T16:24:59.095779Z (0x0000007fa2345fc0) [error] MAIN->Agent::start->Connector::resolve: Host not found (non-authoritative), try again later: Host not found (non-authoritative), try again later … 5 minutes later. I do not see any connection attempts in Agent logs. Changing agent.cfg to IP address instead of hostname, does yield agent reconnect in logs 10/06/2022 11:33:47 AM 2022-10-06T16:33:44.381736Z (0x0000007f8075e0c0) [info] : reconnect: retrying connection (edited) Chris Misztur 11:43 AM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665074592961079> Another observation with 2.0. We sent it invalid SHDR from one of the adapters. There are two adapters in cfg. One sending good SHDR, the other bad SHDR. It seems that the good SHDR is never parsed resulting in data items = UNAVAILABLE for adapter sending good SHDR. 10/06/2022 11:35:36 AM 2022-10-06T16:35:27.208315Z (0x0000007fa163e0c0) [info] Connector::reader->Connector::parseSocketBuffer->Connector::processLine->DataItemMapper.ShdrTokenMapper.operator->DataItemMapper.ShdrTokenMapper.mapTokensToDataItem: Could not find data item: 1.9 1 reply 5 days agoView thread will_s 1:14 PM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080092601969> I’m thinking that the inability to resolve the name results in a fatal error. Probably need some better logs. I don’t think we retry DNS since it assumes it will not appear. I will check the code and see what the behavior is when i get in. 1:15 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080105430969> We can retry if needed 1:15 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080140053489> Do you expect the name to resolve in the future? Are you thinking WINS or some DNS entry 1:15 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080148591759> Will be created 1:15 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080150831729> ? 1:17 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080236462379> So, two options: Retry connection because name will be added Log as fatal connection issue, if no other adapters, exit, otherwise exit adapter 1:19 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080396261709> Just reviewed code. If it does not resolve the address, it fails to start 1:20 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080411474309> Thoughts? Chris Misztur 1:23 PM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080591582359> Retry, this is a docker network, so when the adapter container comes up, the host name will become available. will_s 1:23 PM <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080633909029> Ok, makes sense. I can change the code to fall back to a retry. It will set a timer and retry of fail. 👍 1 1:24 <https://mtconnect.slack.com/archives/C5EB9GES1/p1665080643835889> I’ll put it into the next RC — Reply to this email directly, view it on GitHub <#217>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJ6BCK3FOHQ3MFNHRIV7M3WCXEFDANCNFSM6AAAAAARCUNETU>. You are receiving this because you are subscribed to this thread.