Try not to trigger the GLIBC race condition.

See https://sourceware.org/bugzilla/show_bug.cgi?id=19329 Fixes #62071 Signed-off-by: Pedro Algarvio <[email protected]>
saltstack · May 19, 2022 · 3f07f00 · 3f07f00
1 parent 391abc9
commit 3f07f00
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 8 deletions.
diff --git a/changelog/62071.fixed b/changelog/62071.fixed
@@ -0,0 +1 @@
+Stop trigering the `GLIBC race condition <https://sourceware.org/bugzilla/show_bug.cgi?id=19329>`_ when parallelizing the resolution of the fqnds.
diff --git a/salt/modules/network.py b/salt/modules/network.py
@@ -2,14 +2,14 @@
 Module for gathering and managing network information
 """
 
+import concurrent.futures
 import datetime
 import hashlib
 import logging
 import os
 import re
 import socket
 import time
-from concurrent.futures import ThreadPoolExecutor
 
 import salt.utils.decorators.path
 import salt.utils.functools
@@ -2084,7 +2084,6 @@ def fqdns():
     HOST_NOT_FOUND = 1
     NO_DATA = 4
 
-    grains = {}
     fqdns = set()
 
     def _lookup_fqdn(ip):
@@ -2115,14 +2114,23 @@ def _lookup_fqdn(ip):
     # when the "fqdn" is not defined for certains IP addresses, which was
     # causing that "socket.timeout" was reached multiple times sequentially,
     # blocking execution for several seconds.
-
     try:
-        with ThreadPoolExecutor(8) as pool:
-            for item in pool.map(_lookup_fqdn, addresses):
-                if item:
-                    fqdns.update(item)
+        with concurrent.futures.ThreadPoolExecutor(8) as pool:
+            future_lookups = {
+                pool.submit(_lookup_fqdn, address): address for address in addresses
+            }
+            for future in concurrent.futures.as_completed(future_lookups):
+                address = future_lookups[future]
+                try:
+                    resolved_fqdn = future.result()
+                    if resolved_fqdn:
+                        fqdns.update(resolved_fqdn)
+                except Exception as exc:  # pylint: disable=broad-except
+                    log.error("Failed to resolve address %s: %s", address, exc)
     except Exception as exc:  # pylint: disable=broad-except
-        log.error("Exception while creating a ThreadPool for resolving FQDNs: %s", exc)
+        log.error(
+            "Exception while creating a ThreadPoolExecutor for resolving FQDNs: %s", exc
+        )
 
     elapsed = time.time() - start
     log.debug("Elapsed time getting FQDNs: %s seconds", elapsed)