3.8-8 update is now available. We’ve made a bunch of improvements. To name the top two iteams, we’ve added support for UPS(thanks for @phillxnet) and Shadow copy support for Windows Samba clients.
Test coverage also improved, thanks to @mchakravartula. Kernel is updated to 4.2.2 and btrfs-progs to 4.2.1. Finally, I’ve also managed to implement Stable and Testing updates channels going forward. As usual, more details in the blog post.
How does the shadow copy feature work? Is it just available by default on samba shares that have snapshots, or does it need to be configured?
maybe for @phillxnet - as I mentioned here, I have a cyberpower ups, with their rpm package installed to manage it, I notice the NUT database mentions a generic driver for it. Any recommendations of which I should use: my current solution, or this NUT one? Also, FWIW, on the setup UI, I always get a little nervous when I’m forced to create accounts and password. Is there not some way to bury this for the average user (maybe as general rule, if a user doesn’t really need to think about permissions like these, put it in an “advanced settings” area)?
PS - no problems updating. I’m not sure this is new, though, but I seem to have problems rebooting from within Rockstor. Each time I reboot from within Rockstor, it gets stuck, and I need to reboot again from the hardware.
So you go to the samba area on the “file sharing” sidebar, and enable the feature.
Observation: the configuration UI is a little confusing. I guess the idea is you create a scheduled snapshot on the share, assign it a prefix, and then use the exact same prefix on the shadow copy config.
Reboot and Shutdown from the Web-UI are scheduled after a minute. The UI freezes, but it seems like nothing happens to the actual system for couple of minutes. I think it’s also possible for reboot/shutdown to wait for certain IO operations to terminate.
@bdarcus Yes funny you should say about the user / pass settings in nut config. @suman and I had discussed hiding / auto dealing with these but I’m afraid I just didn’t get around to it and so they are left open. In the standalone mode they could be auto generated as they are only for intra nut function. Don’t use a regular user just follow the mouse over suggestions and go with some nonsense user and pass. I will try and sort this in a future pull request.
As for if you should try out the built in nut config or stick with what’s working for you now then that’s down really to how curious you are. As you say the nut driver for your particular UPS model is a generic one and apparently as per the NUT Hardware compatibility list it only shows as 2 our of 5 stars due to no manufacturer participation but on the other hand there are successful reports of older NUT releases retrieving info successfully, ie heres a couple of CP1000PFCLCD info dumps via “upsc ups” commands.
If you do decide to take the plunge then remember to first “yum remove” the rpm package you put in before configuring and enabling the nut service, you might also have to reboot after the yum remove, just to be safe.
As it stands there is little immediate advantage in using the built in nut, except maybe email notifications, but that should change over time. Although I would hazard a guess that if it does work for you it will be easier to configure. Also if you use netserver mode (but get standalone mode working first) you can use fancy desktop apps to monitor you ups over the network. If you also power other machines on the network and the network itself from this UPS you could even have them all shutdown automatically followed shortly after by Rockstor by running nut clients on them and Rockstor’s nut in netserver mode as master. You might like that; you have 600W to play with there so keep an eye on load.
So short of it is if this is critical leave as is as it’s working but otherwise it would be great to know how you get on. Though maybe in a specific new thread so we don’t clutter this one up.