svn post-commit Temporary failure in name resolution

Question

First, I have svn configured to be accessible via apache DAV.

Then I have a post-commit hook that tell jenkins continuous integration server that the repository has changed via a simple wget command (and make it build) :

/opt/local/bin/wget (many-options) https://jenkins.domain.tld/subversion/${UUID}/notifyCommit?rev=$REV >> $LOGFILE

The script checks wget return code, and then echos an error message and exit with 1 (so the svn client knows the commit failed).

The LOGFILE contains after a failure :

--2012-04-04 10:56:54--  https://jenkins.domain.tld/subversion/4d729FFF-FFFF-4fb8-b8b5-be083cf6fe96/notifyCommit?rev=302
Résolution de jenkins.domain.tld (jenkins.domain.tld)... échec: Temporary failure in name resolution.
wget : impossible de résoudre l'adresse de l'hôte «jenkins.domain.tld»

The host jenkins.domain.tld has been added to /etc/hosts to be sure that resolution can't fail, but I'm still having the issue.

Of course, I have checked the post-commit command from CLI, and it works.

I've added this line to the post-commit script :

/usr/bin/host jenkins.domain.tld >> $LOGFILE

And resolution works just fine... maybe a wget issue.

Just replaced wget call with a curl equivalent, same issue :

curl: (6) Couldn't resolve host 'jenkins.domain.tld'

Answer 1

Ok, not found the issue origin, but a solution (pretty ugly) :

Update: I solved it meanwhile by fully stopping apache, waiting for a moment, then starting it again. But that doesn't explain the root cause for this behaviour (and thus there's no satisfying solution). So I leave this open.

From : https://serverfault.com/questions/129772/php-gethostbyname-suddenly-no-longer-resolves-names-to-ips-when-run-in-apache

Answer 2

After upgrading to Ubuntu 14.04 LTS I had the same issue (not sure whether the upgrade was really related). However, I could solve it by adding explicit timeout configuration to wget in my post-commit hook:

wget \
   --timeout=10 \
   --dns-timeout=10 \
...