Univention Bugzilla – Bug 37817
Test 43_proxy.04_http_proxy_ntlm_auth_performance_check.test failed in jenkins
Last modified: 2023-03-25 06:43:31 CET
Happens in 3.2->4.0 Update App Tests: Fehlermeldung Test failed Standard Ausgabe (STDOUT) Create squid/ntlmauth File: /etc/squid3/squid.conf Restarting Squid HTTP Proxy 3.x: squid3 Waiting.....................done. . Test took 106 seconds Cleanup ... Unsetting squid/ntlmauth File: /etc/squid3/squid.conf Restarting Squid HTTP Proxy 3.x: squid3 Waiting.....................done. . ### FAIL ### test took to long (max 60) ### ### Standard Fehler (STDERR) W: The config registry variable 'squid/basicauth' does not exist W: The config registry variable 'squid/krb5auth' does not exist
Also happens in 4.0 App Auto Tests. Quite frequently.
Happens in 4.0-2 all the time, new Output: (2015-06-26 07:14:49.772201)W: The config registry variable 'squid/basicauth' does not exist (2015-06-26 07:14:49.772295)W: The config registry variable 'squid/krb5auth' does not exist [2015-06-26 07:14:49.800300]Create squid/ntlmauth [2015-06-26 07:14:49.926012]File: /etc/squid3/squid.conf [2015-06-26 07:14:52.018971]Restarting Squid HTTP Proxy 3.x: squid3 failed! [2015-06-26 07:14:55.028353]<(7, "couldn't connect to host") [2015-06-26 07:14:55.029948]>### FAIL ### [2015-06-26 07:14:55.029965]job 29421 failed with status 1 [2015-06-26 07:14:55.029974]### ### [2015-06-26 07:14:55.045754]Unsetting squid/ntlmauth [2015-06-26 07:14:55.176205]File: /etc/squid3/squid.conf [2015-06-26 07:14:57.301451]Restarting Squid HTTP Proxy 3.x: squid3.
{S3,S4}-Master sind 10 runs: <http://jenkins.knut.univention.de:8080/job/UCS-4.0/job/UCS-4.0-2/job/Autotest%20MultiEnv/52/SambaVersion=s4,Systemrolle=master/testReport/43_proxy/04_http_proxy_ntlm_auth_performance_check/test/> <http://jenkins.knut.univention.de:8080/job/UCS-4.0/job/UCS-4.0-2/job/Autotest%20MultiEnv/52/SambaVersion=s3,Systemrolle=master/testReport/43_proxy/04_http_proxy_ntlm_auth_performance_check/test/> I was able to repduce it locally on my S4-Slave after running the test for ~3 hours in an endless loop: # cat /var/log/squid3/cache.log ... 2015/07/02 15:17:37| WARNING: All ntlmauthenticator processes are busy. 2015/07/02 15:17:37| WARNING: 10 pending requests queued 2015/07/02 15:17:37| Consider increasing the number of ntlmauthenticator processes in your config file. 2015/07/02 15:17:37| storeDirWriteCleanLogs: Starting... 2015/07/02 15:17:37| WARNING: Closing open FD 36 2015/07/02 15:17:37| Finished. Wrote 2 entries. 2015/07/02 15:17:37| Took 0.00 seconds (4210.53 entries/sec). FATAL: Too many queued ntlmauthenticator requests Squid Cache (Version 3.1.20): Terminated abnormally. CPU Usage: 1.564 seconds = 1.004 user + 0.560 sys Maximum Resident Size: 86112 KB Page faults with physical i/o: 0 Memory usage for squid via mallinfo(): total space in arena: 3440 KB Ordinary blocks: 3354 KB 24 blks Small blocks: 0 KB 1 blks Holding blocks: 36124 KB 9 blks Free Small blocks: 0 KB Free Ordinary blocks: 85 KB Total in use: 39478 KB 1148% Total free: 85 KB 2% Traceback (most recent call last): File "/usr/lib/squid3/squid_ldap_ntlm_auth", line 689, in <module> sys.stdout.write(answer + "\n") IOError: [Errno 32] Broken pipe ... Squid is still overloaded by too many parallel NTLM authentication requests. In EC2 the ressources seem to be much more limited, so the test seems to fail more often there. Disable the test as the result is not reliable enough?
*** This bug has been marked as a duplicate of bug 37931 ***
For this bug is no separate QA needed.