Segmentation fault

I have a ceph cluster.
Use "ceph osd crush dump > map.json" to get the crushmap about my ceph cluster.
And then, run "crush analyze --rule replicated --crushmap map.json", it occurs a segmentation fault.
The log display in dmesg is "[12471.824472] crush[27369]: segfault at ffffffffffffffff ip 00000000004539b9 sp 00007ffe8a036370 error 5 in python2.7[400000+2bb000]".

Do you have some comments about this? If you have, i will appreciate you very much.
The attached file is the crushmap file about the ceph cluster.
[mao.zip](https://github.com/ceph/python-crush/files/1158228/mao.zip)

首先解压附加的 mao.zip 文件，并使用 `ceph osd crush dump > map.json`，然后执行 `crush analyze --rule replicated --crushmap map.json` 来重现命令序列。检查由此产生的 Python 2.7 段错误，并在附加的 crushmap 被无崩溃地分析后确认完成。

由索引模型根据 Issue 内容生成。