# Need advise on define inventory-specific variables **URL:** \ **Category:** Ansible Project\ **Created:** [July 7, 2014, 4:11am UTC](https://forum.ansible.com/t/need-advise-on-define-inventory-specific-variables/15616 "2014-07-07T04:11:29Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![Chengbing\_Liu](https://avatars.discourse-cdn.com/v4/letter/c/b2d939/32.png) [@Chengbing\_Liu](https://forum.ansible.com/u/Chengbing_Liu)\ **Post date:** [July 7, 2014, 4:11am UTC](https://forum.ansible.com/t/need-advise-on-define-inventory-specific-variables/15616/1 "2014-07-07T04:11:29Z")
I’m using Ansible to deploy and maintain Hadoop clusters. I used to define variables in group\_vars/all as shown below: > hadoop: > > # Variables for core-site.xml > > nameservice: hadoop-namenode > > # Variables for hdfs-site.xml > > dfs\_blocksize: 256m > > # Variables for mapred-site.xml > > mapred\_job\_tracker: hadoop-jobtracker:8021 However, I have different inventory files, which makes it prone to mistakes. I can define variables using [all:vars] in inventory files: > [all:vars] > hadoop\_nameservice=hadoop-namenode How can I define inventory-specific variables, while still preserving the structured definition as in group\_vars/all? Thanks a lot! ---
**Author:** ![Michael\_DeHaan1](https://avatars.discourse-cdn.com/v4/letter/m/5e9695/32.png) [@Michael\_DeHaan1](https://forum.ansible.com/u/Michael_DeHaan1)\ **Post date:** [July 7, 2014, 7:29pm UTC](https://forum.ansible.com/t/need-advise-on-define-inventory-specific-variables/15616/2 "2014-07-07T19:29:08Z")
Perhaps I’m not understanding the question, but: group\_vars/all group\_vars/groupname1 group\_vars/groupname2 etc This can be used with multiple inventory files if group\_vars is a directory in the directory with both of them, which should be what you need. Let us know if this doesn’t make sense and perhaps I can elaborate further. Thanks! ---
**Author:** ![Chengbing\_Liu](https://avatars.discourse-cdn.com/v4/letter/c/b2d939/32.png) [@Chengbing\_Liu](https://forum.ansible.com/u/Chengbing_Liu)\ **Post date:** [July 8, 2014, 9:41am UTC](https://forum.ansible.com/t/need-advise-on-define-inventory-specific-variables/15616/3 "2014-07-08T09:41:05Z")
Sorry, the question was not stated clearly. I have different inventory files corresponding to different clusters, e.g.: hosts\_cluster1 > [NameNode] > host1 hosts\_cluster2 > [NameNode] > host2 They have the same group names, therefore I decide to move the variable definitions from group\_vars to the all:vars section of inventory file. However, I cannot define structured variables like before in group\_vars. Any advice? 在 2014年7月8日星期二UTC+8上午3时29分14秒，Michael DeHaan写道： > Perhaps I’m not understanding the question, but: > > group\_vars/all > group\_vars/groupname1 > > group\_vars/groupname2 > > etc > > This can be used with multiple inventory files if group\_vars is a directory in the directory with both of them, which should be what you need. > > Let us know if this doesn’t make sense and perhaps I can elaborate further. > > Thanks! > > > I’m using Ansible to deploy and maintain Hadoop clusters. > > > > I used to define variables in group\_vars/all as shown below: > > > > > hadoop: > > > > > > # Variables for core-site.xml > > > > > > nameservice: hadoop-namenode > > > > > > # Variables for hdfs-site.xml > > > > > > dfs\_blocksize: 256m > > > > > > # Variables for mapred-site.xml > > > > > > mapred\_job\_tracker: hadoop-jobtracker:8021 > > > > However, I have different inventory files, which makes it prone to mistakes. > > > > I can define variables using [all:vars] in inventory files: > > > > > [all:vars] > > > hadoop\_nameservice=hadoop-namenode > > > > How can I define inventory-specific variables, while still preserving the structured definition as in group\_vars/all? > > > > Thanks a lot! > > > > – > > You received this message because you are subscribed to the Google Groups “Ansible Project” group. > > To unsubscribe from this group and stop receiving emails from it, send an email to [ansible-proje...@googlegroups.com](mailto:ansible-proje...@googlegroups.com). > > To post to this group, send email to [ansible...@googlegroups.com](mailto:ansible...@googlegroups.com). > > To view this discussion on the web visit [https://groups.google.com/d/msgid/ansible-project/a62a724b-d9db-4bb0-980d-91efff5c869b%40googlegroups.com](https://groups.google.com/d/msgid/ansible-project/a62a724b-d9db-4bb0-980d-91efff5c869b%40googlegroups.com?utm_medium=email&utm_source=footer). > > For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout). 在 2014年7月8日星期二UTC+8上午3时29分14秒，Michael DeHaan写道： > Perhaps I’m not understanding the question, but: > > group\_vars/all > group\_vars/groupname1 > > group\_vars/groupname2 > > etc > > This can be used with multiple inventory files if group\_vars is a directory in the directory with both of them, which should be what you need. > > Let us know if this doesn’t make sense and perhaps I can elaborate further. > > Thanks! > > > I’m using Ansible to deploy and maintain Hadoop clusters. > > > > I used to define variables in group\_vars/all as shown below: > > > > > hadoop: > > > > > > # Variables for core-site.xml > > > > > > nameservice: hadoop-namenode > > > > > > # Variables for hdfs-site.xml > > > > > > dfs\_blocksize: 256m > > > > > > # Variables for mapred-site.xml > > > > > > mapred\_job\_tracker: hadoop-jobtracker:8021 > > > > However, I have different inventory files, which makes it prone to mistakes. > > > > I can define variables using [all:vars] in inventory files: > > > > > [all:vars] > > > hadoop\_nameservice=hadoop-namenode > > > > How can I define inventory-specific variables, while still preserving the structured definition as in group\_vars/all? > > > > Thanks a lot! > > > > – > > You received this message because you are subscribed to the Google Groups “Ansible Project” group. > > To unsubscribe from this group and stop receiving emails from it, send an email to [ansible-proje...@googlegroups.com](mailto:ansible-proje...@googlegroups.com). > > To post to this group, send email to [ansible...@googlegroups.com](mailto:ansible...@googlegroups.com). > > To view this discussion on the web visit [https://groups.google.com/d/msgid/ansible-project/a62a724b-d9db-4bb0-980d-91efff5c869b%40googlegroups.com](https://groups.google.com/d/msgid/ansible-project/a62a724b-d9db-4bb0-980d-91efff5c869b%40googlegroups.com?utm_medium=email&utm_source=footer). > > For more options, visit [https://groups.google.com/d/optout](https://groups.google.com/d/optout). 在 2014年7月8日星期二UTC+8上午3时29分14秒，Michael DeHaan写道： ---
**Author:** ![Michael\_DeHaan1](https://avatars.discourse-cdn.com/v4/letter/m/5e9695/32.png) [@Michael\_DeHaan1](https://forum.ansible.com/u/Michael_DeHaan1)\ **Post date:** [July 8, 2014, 11:57am UTC](https://forum.ansible.com/t/need-advise-on-define-inventory-specific-variables/15616/4 "2014-07-08T11:57:14Z")
“They have the same group names, therefore I decide to move the variable definitions from group\_vars to the all:vars section of inventory file. However, I cannot define structured variables like before in group\_vars.” I would consider different group names for each, and a parent group that contains them called NameNode or similar. This way they could be selected by the same group but have different values for variables.