zhaojs comments

Results 5 comments of


                                            zhaojs

pdns recursor crash on startup，The probability of crash is relatively high

**bt:** #8 0x00005555559419ef in SyncRes::shuffleInSpeedOrder (this=this@entry=0x7fffd405cec0, tnameservers=std::unordered_map with 13 elements = {...}, prefix="", auth=...) at syncres.cc:1808 #9 0x000055555591b62b in SyncRes::doResolveAt (this=this@entry=0x7fffd405cec0, nameservers=..., auth=..., flawedNSSet=, flawedNSSet@entry=false, qname=..., qtype=..., ret=..., depth=, beenthere=...,...

pdns recursor crash on startup，The probability of crash is relatively high

I also looked at the logic, maybe my analysis is wrong, sorry

When the recursor is configured to use the TCP protocol for outbound query, TCP fd leakage problem will occur

I'm forcing the use of bare TCP protocol, not DOT. When there is no second parsing of the same request, the FD status CLOSE_WAIT will always exist. The log I...

When the recursor is configured to use the TCP protocol for outbound query, TCP fd leakage problem will occur

Yes, **MTasker** holds a reference to the smart pointer. When the worker thread makes a request, the cleanup is quite timely. If there is no parsing request, fd will be...

When cname itself is found in the cname cache, an exception is thrown directly

> Does [#14822](https://github.com/PowerDNS/pdns/pull/14822) and [#15396](https://github.com/PowerDNS/pdns/pull/15396) have any influence on this behaviour? It would be nice to test against a more recent version. The problem remains unresolved. My understanding is that...