Skip to content

Revert "DNS: Fix lack of nameserver failover in low use circumstances… #8663

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
bneradt merged 1 commit into from
Feb 11, 2022
Merged

Revert "DNS: Fix lack of nameserver failover in low use circumstances… #8663

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 7 additions & 21 deletions iocore/dns/DNS.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -524,12 +524,12 @@ DNSHandler::open_con(sockaddr const *target, bool failed, int icon, bool over_tc
}
return false;
} else {
ns_down[icon] = 0;
if (cur_con.eio.start(pd, &cur_con, EVENTIO_READ) < 0) {
Error("[iocore_dns] open_con: Failed to add %d server to epoll list\n", icon);
} else {
cur_con.num = icon;
ns_down[icon] = 0;
Debug("dns", "opening connection %s on fd %d SUCCEEDED for %d", ip_text, cur_con.fd, icon);
cur_con.num = icon;
Debug("dns", "opening connection %s SUCCEEDED for %d", ip_text, icon);
}
ret = true;
}
Expand Down Expand Up @@ -729,15 +729,6 @@ void
DNSHandler::failover()
{
Debug("dns", "failover: initiating failover attempt, current name_server=%d", name_server);
if (!ns_down[name_server]) {
ip_text_buffer buff;
// mark this nameserver as down
Debug("dns", "failover: Marking nameserver %d as down", name_server);
ns_down[name_server] = 1;
Warning("connection to DNS server %s lost, marking as down",
ats_ip_ntop(&m_res->nsaddr_list[name_server].sa, buff, sizeof(buff)));
}

// no hope, if we have only one server
if (m_res->nscount > 1) {
ip_text_buffer buff1, buff2;
Expand Down Expand Up @@ -775,8 +766,6 @@ DNSHandler::failover()
ip_text_buffer buff;
Warning("failover: connection to DNS server %s lost, retrying", ats_ip_ntop(&ip.sa, buff, sizeof(buff)));
}
// Make sure retries are done even if no more requests.
this_ethread()->schedule_in(this, DNS_PRIMARY_RETRY_PERIOD);
}

/** Mark one of the nameservers as down. */
Expand All @@ -790,8 +779,6 @@ DNSHandler::rr_failure(int ndx)
Debug("dns", "rr_failure: Marking nameserver %d as down", ndx);
ns_down[ndx] = 1;
Warning("connection to DNS server %s lost, marking as down", ats_ip_ntop(&m_res->nsaddr_list[ndx].sa, buff, sizeof(buff)));
// Make sure retries are done even if no more requests.
this_ethread()->schedule_in(this, DNS_PRIMARY_RETRY_PERIOD);
}

int nscount = m_res->nscount;
Expand Down Expand Up @@ -984,11 +971,6 @@ DNSHandler::check_and_reset_tcp_conn()
int
DNSHandler::mainEvent(int event, Event *e)
{
// If this was a scheduled retry event, clear the associated flag.
if (e && e->cookie == RETRY_COOKIE) {
this->nameserver_retry_in_flight_p = false;
}

recv_dns(event, e);
if (dns_ns_rr) {
if (DNS_CONN_MODE::TCP_RETRY == dns_conn_mode) {
Expand Down Expand Up @@ -1035,6 +1017,10 @@ DNSHandler::mainEvent(int event, Event *e)
write_dns(this);
}

if (std::any_of(ns_down, ns_down + n_con, [](int f) { return f != 0; })) {
this_ethread()->schedule_at(this, DNS_PRIMARY_RETRY_PERIOD);
}

return EVENT_CONT;
}

Expand Down
23 changes: 8 additions & 15 deletions iocore/dns/P_DNSProcessor.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -169,16 +169,9 @@ struct DNSHandler : public Continuation {
DNSConnection udpcon[MAX_NAMED];
Queue<DNSEntry> entries;
Queue<DNSConnection> triggered;
int in_flight = 0;
int name_server = 0;
int in_write_dns = 0;
/// Rate limiter for down nameserver retries.
/// Don't schedule another if there is already one in flight.
std::atomic<bool> nameserver_retry_in_flight_p{false};
/// Marker for event cookie to indicate it's a nameserver retry event.
/// @note Can't be @c constexpr because of the cast.
static inline void *const RETRY_COOKIE{reinterpret_cast<void *>(0x2)};

int in_flight = 0;
int name_server = 0;
int in_write_dns = 0;
HostEnt *hostent_cache = nullptr;

int ns_down[MAX_NAMED];
Expand Down Expand Up @@ -219,16 +212,16 @@ struct DNSHandler : public Continuation {
(ink_hrtime)HRTIME_SECONDS(dns_failover_period));
Debug("dns", "\tdelta time is %" PRId64 "", (Thread::get_hrtime() - crossed_failover_number[i]));
}
return ns_down[i] || (crossed_failover_number[i] &&
((Thread::get_hrtime() - crossed_failover_number[i]) > HRTIME_SECONDS(dns_failover_period)));
return (crossed_failover_number[i] &&
((Thread::get_hrtime() - crossed_failover_number[i]) > HRTIME_SECONDS(dns_failover_period)));
}

bool
failover_soon(int i)
{
return ns_down[i] || (crossed_failover_number[i] &&
((Thread::get_hrtime() - crossed_failover_number[i]) >
(HRTIME_SECONDS(dns_failover_try_period + failover_soon_number[i] * FAILOVER_SOON_RETRY))));
return (crossed_failover_number[i] &&
((Thread::get_hrtime() - crossed_failover_number[i]) >
(HRTIME_SECONDS(dns_failover_try_period + failover_soon_number[i] * FAILOVER_SOON_RETRY))));
}

void recv_dns(int event, Event *e);
Expand Down