Project

General

Profile

Tâche #11218

Distribution EOLE - Scénario #11062: Publier la mise à jour 2.3.16 Stable

Problèmes de plantage du backend Zéphir suite à mise à jour 2.3.16 RC

Added by Bruno Boiget about 9 years ago. Updated about 9 years ago.

Status:
Reporté
Priority:
Normal
Assigned To:
Start date:
04/15/2015
Due date:
% Done:

0%

Estimated time:
10.00 h
Spent time:
Remaining (hours):
0.0

Description

Suite à des tests sur les derniers paquets candidats, des remontées ont été faites concernant des plantages fréquents du backend Zéphir.

différents symptomes ont été remontés :

- segfault du processus avec des erreurs de ce type (constatés à Lyon et sur les serveurs Zéphir du Medde). En partciulier, certains segfault semble se produire sur la librairie etree.so

[824254.914007] twistd[3747]: segfault at 4 ip 0810d0a1 sp a1454410 error 6 in python2.6[8048000+1e1000] (besançon)
[  164.798183] twistd[3961]: segfault at 7200000061 ip 00000000004d71d0 sp 00007f1cfeedf330 error 4 in python2.6[400000+21d000] (thierry)
[ 1087.216732] twistd[12042]: segfault at 8 ip 00000000004d6a0e sp 00007f7b6b2bc808 error 6 in python2.6[400000+21d000]
karim[ac-lyon]> [968853.042550] twistd[3460]: segfault at 8 ip 00000000004d6a0e sp 00007f0215630598 error 6 in python2.6[400000+21d000]
<karim[ac-lyon]> [1149075.256246] twistd[15499]: segfault at 7500000041 ip 0000000000456dea sp 00007f10824d1600 error 4 in python2.6[400000+21d000]
<karim[ac-lyon]> [1749261.295040] twistd[24456]: segfault at 7f584dd19a70 ip 00007f581c7cda6d sp 00007f58013ad5a0 error 4 in libc-2.11.1.so[7f581c755000+17f000]
<karim[ac-lyon]> [1761839.561921] twistd[9513]: segfault at 7200000065 ip 00007ff775d2c56e sp 00007ff75b1f9800 error 4 in etree.so[7ff775d02000+108000]
<karim[ac-lyon]> [2352852.289447] twistd[9336]: segfault at 8 ip 00000000004d6a0e sp 00007f852e7f9808 error 6 in python2.6[400000+21d000]
<karim[ac-lyon]> [2366609.420815] twistd[15071]: segfault at 74000000f4 ip 0000000000456dea sp 00007fdc177fafc0 error 4 in python2.6[400000+21d000]
<karim[ac-lyon]> [2366843.524214] twistd[13503]: segfault at 8 ip 00000000004d6a0e sp 00007f8fa5ff8808 error 6 in python2.6[400000+21d000]

- problème à priori différent remonté par Versailles (Arnaud Bougeard) :

Traceback (most recent call last):
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 1371, in dataReceived
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    finishCallback(data[contentLength:])
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 1585, in _finishRequestBody
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    self.allContentReceived()
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 1641, in allContentReceived
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    req.requestReceived(command, path, version)
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 807, in requestReceived
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    self.process()
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011--- <exception caught here> ---
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/server.py", line 125, in process
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    self.render(resrc)
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/twisted/web/server.py", line 132, in render
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    body = resrc.render(self)
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011  File "/usr/lib/python2.6/dist-packages/zephir/backend/xmlrpceole.py", line 137, in render
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011    cx = PgSQL.connect(database=config.DB_NAME,user=config.DB_USER,password=config.DB_PASSWD)
Apr 10 14:25:21 zephir zephir_backend: [HTTPChannel,333894,213.41.210.90] #011psycopg2.OperationalError: could not create socket: Too many open files

Apr 10 14:25:23 zephir zephir_backend: [twisted.web.server.Site] Could not accept new connection (EMFILE)
Apr 10 14:25:23 zephir zephir_backend: last message repeated 9 times
Apr 10 14:25:23 zephir zephir_backend: [twisted.web.server.Site] Could not accept new connection (EMFILE)
Apr 10 14:25:23 zephir zephir_backend: last message repeated 467 times

Related issues

Related to conf-zephir - Tâche #11211: segfault Zephir Fermé 04/15/2015

Associated revisions

Revision 6e1d6323 (diff)
Added by Bruno Boiget about 9 years ago

Correction de la fermeture de connexions à la base en cas d'erreurs.

ref #11218 @1h

History

#1 Updated by Bruno Boiget about 9 years ago

  • Remaining (hours) changed from 10.0 to 6.0

#2 Updated by Bruno Boiget about 9 years ago

  • Status changed from Nouveau to En cours
  • Remaining (hours) changed from 6.0 to 2.0

Dans l'état actuel, le problème n'est pas résolu mais un contournement semble possible en désactivant l'utilisation des threads dans le backend Zéphir

-> gen_config (mode expert) -> personnalisation -> passer "Activer l'utilisation de threads" à "non"

Cela semble avoir stabilisé la situation sur un serveur Zéphir du MEDDE.

Voir avec Bourritux si le serveur continue de fonctionner demain, et faire une annonce sur la liste Zéphir (éventuellement aussi en réponse du mail sur la liste Amon) pour recommander temporairement cette solution.

#3 Updated by Joël Cuissinat about 9 years ago

  • Status changed from En cours to Reporté
  • Parent task set to #11062
  • Remaining (hours) changed from 2.0 to 0.0

Also available in: Atom PDF