We've got a series of scripts we run on our normal mysql servers that do a flush tables with writelock prior to starting VMWare snapshot and then unlock tables after (it's about 7 seconds of writelock). Trying to do the same on HipChat server is proving difficult due to mysql rules and general appliance-ization of this "server."
Given that Atlassian insists on only distributing HipChat as a virtualized appliance (making HipChat an awful martian in pretty much anyone's environment, by the way), what is the current guidance on getting consistent VM snapshots?
Hi Mike.
Here’s the HipChat documentation for making backups: Back up and restore HipChat Server
Hopefully that gives you some useful info.
Hi John,
Thanks for your response. I've read the page you linked to and unfortunately that does not address my question about getting an application-consistent backup of the mysql database on the HipChat server, only crash-consistent. Given the amount of write activity on a typical environment, there's a good chance of this leading to a messy restore.
So I must repeat my previous question: how do I get a consistent database backup from a VM snapshot? We do this on all our other mysql databases very easily, but then again they are VMs built against our shop standard and thus manageable.
Having HipChat as a regular install on our Linux hosts would eliminate the need for this question and many others to come. I've read the official explanation of why HipChat is only being distributed as a VM and not an installable application, and frankly it does not come close to passing the smell test. I read "it's too much work!" in the subtext. Please let us know when Atlassian plans to fix this nagging issue.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.