[Varnish] #1528: Assert error in hcb_insert(), hash/hash_critbit.c line 216: Condition((y)->magic == 0x125c4bd2) not true.

[Varnish]

#1528: Assert error in hcb_insert(), hash/hash_critbit.c line 216:
Condition((y)->magic == 0x125c4bd2) not true.
----------------------+--------------------
 Reporter:  joakim    |       Owner:
     Type:  defect    |      Status:  new
 Priority:  normal    |   Milestone:
Component:  varnishd  |     Version:  4.0.0
 Severity:  normal    |  Resolution:
 Keywords:            |
----------------------+--------------------

Comment (by joakim):

 Ticket #1527 is using the build provided by repo.varnish-cache.org^1^ and,
 yeah, it seems to be using glibc malloc. Tickets #1527, #1528 and #1529
 are using the deb created by running dpkg-buildpackage -b on the master
 branch in this git repo: http://anonscm.debian.org/gitweb/?p=pkg-varnish
 /pkg-varnish.git;a=shortlog and it is using jemalloc.

 - Low load generally (avg ~2.5 with ~500-700 connections on a 24 core Xeon
 E5 machine).
 - A few hours between failures, sometimes longer sometimes shorter.
 Longest uptime I've seen is just under a day.
 - Haven't really been able to see any patterns, though expunging of
 expired objects usually go on when it does crash.

 The machines have been running Varnish 3.0.5 for a long time without a
 hick-up on the same hardware.

 I'll sanitize my VCLs and attach them too when I get the chance.

 ---

 ^1^ http://repo.varnish-
 cache.org/debian/pool/varnish-4.0/v/varnish/varnish_4.0.0-1~wheezy_amd64.deb

-- 
Ticket URL: <https://www.varnish-cache.org/trac/ticket/1528#comment:2>
Varnish <https://varnish-cache.org/>
The Varnish HTTP Accelerator