問題

The export from Hipchat Server or Data Center fails when encrypting the archive. The following appears in /var/log/hipchat/hcapp.log:

2018-07-04T15:36:47.246068+00:00 nr-hipchat scissortail-worker:18159:INFO GID: 1 JID: 29f75629-e6f6-4e99-a575-56523745ab9e - Couldn't encrypt the file: Command '['openssl', 'aes-256-cbc', '-in', '/file_store/scissortail/tmpxjpo7dsx/hipchat-2018-07-04_14-42-44.tar.gz', '-out', '/file_store/scissortail/tmpxjpo7dsx/hipchat-2018-07-04_14-42-44.tar.gz.aes', '-pass', 'env:ada18de7e6854a8d93d9f864e58dfa32']' returned non-zero exit status 1

診断

• The issue can occur in Server and Data Center environments.
• The issue is more likely to occur if you include attachments in the export. 

原因

Scissortail exports the data fine but fails to encrypt the archive. This is due to the lack of disk space in the server/DC node. 

回避策

Scissortail uses an intermediate directory to encrypt the archive before doing the cleanup. This directory is defined by Chef in /hipchat-scm/chef-repo/cookbooks/scissortail/attributes/default.rb and points to /file_store/scissortail by default:

default[:scissortail][:working_dir] = "/file_store/scissortail"

Server

• In Server, the folder can be changed to point to a location in another disk. 
• Alternatively, you can increase disk space.

Data Center

• In DC, this can be changed to point to /file_store/shared/scissortail which is under the NFS share. 
• We have a feature to default the work directory to a location under the NFS share. Watch and up-vote  HCPUB-2195 - Getting issue details... STATUS

After the change

1. Run Chef-Solo as root: 

   /opt/atlassian/hipchat/sbin/cs

2. Restart Scissortail, it will create the new temp directory automatically: 

   /etc/init.d/scissortail-web-0 restart && /etc/init.d/scissortail-worker-0 restart