RouterOS 7.20 running OK for 4 months, then all /tool fetch url="https://hc-ping.com/… keep-result=no timeout= run by scheduler and netwatch scripts stopped working: “download from hc-ping failed: ssl: internal error (6)”
Test:/tool fetch url=``https://www.google.com`` keep-result=no failed: “failure: maximum connection count reached”
Connection count was only 73.
Reboot fixed. What may have caused? Advice on how to prevent?
Thanks!
You may want to add some logging to see where the problem may lay. Something like:
/system/logging/add topics=fetch action=memory
which may be a lot, so can disable after troubleshooting
But are you sure it's not upstream internet that may be the cause? For example, the failure message could from from a HTTP error message returned. Why logging may help narrow down cause of the error.
And, on example, perhaps a cut-and-paste error. But `` quotes are not valid.
Internet is fine for everything except /tool fetch, which stops working completely until router rebooted. ChatGPT and Gemini tell me that duration= should be used instead of idle-timeout= because idle-timeout= doesn’t actually time out in some cases, eventually causing exhaustion of the SSL connection pool. Does that make sense? Would explain why it took four months for the failure to occur.
Have you tried not setting any idle-timeout= and duration=?
Is there a reason you think you need them?
Single /tool fetch to ping healthchecks.io wasn’t reliable. I added idle-timeout= and ran /tool fetch in a retry loop, which worked OK until this new problem.