Dev log for upcoming 3.8-11 [Latest: 3.8-10.18]

suman · December 30, 2015, 3:10am

Turns out we are not done for the year just yet. We just released 3.8-10.09! The following two issues were closed. Test away!

github.com/rockstor/rockstor-core

Failed to delete the Share(Images). Error from the OS: 'NoneType' object has no attribute 'name'

opened 08:55PM - 06 Oct 15 UTC

closed 09:25PM - 29 Dec 15 UTC

schakrava

Critical

In my installation I had a single harddrive with a pool on it an a share over th…e whole drive. I removed this drive because of failed sectors and did not replace it by another. Now I am not able to remove the share or the pool in the UI. If I go to delete the share, I get the following: Failed to delete the Share(Images). Error from the OS: 'NoneType' object has no attribute 'name' If I go to delete the pool, I get this: Pool(pool1) is not empty. Delete is not allowed until all shares in the pool are deleted

suman · December 31, 2015, 7:15pm

3.8-10.10 is now out! We’ve managed to close two more issues(thanks @phillxnet) and are inching forward to -11 release. There are many important changes to UI-internals that are brewing in this pull request, thanks to @ganti_priya. I really want these changes to go into -11 release, but we need to test a lot more. A decent amount of Rock-On related work is also done/underway, but I think there won’t be any UI changes this time around because I don’t want to mix them up with the ones coming from priya.

Anyways, this I am sure is the last update for 2015. Happy new year! Drink responsibly

github.com/rockstor/rockstor-core

enhance scan_disks to handle md block devices

opened 08:19PM - 22 Dec 15 UTC

closed 07:12PM - 31 Dec 15 UTC

phillxnet

Although the build in fail over to use udev rather than lsblk to find an md bloc…k device's serial number succeeds the fact that it appears twice in lsblk's output (in the case of a 2 device md) then triggers the duplication removal logic and the real udev rule applied serial number successfully retrieved on the first pass is overwritten with an auto generated fake-serial on the second. This in turn causes the "Disk Serial number not legitimate" warning even when a valid serial is available (applied via udev rule). N.B. this has only been proved on non root md devices but is strongly suspected to apply to the system device also which is the only likely supported use for md devices. Please update the following forum post on progress with this issue:- http://forum.rockstor.com/t/issues-cant-wipe-some-drives-no-smart-diagnostics-disk-serial-numbers-not-legitamate/820

suman · December 31, 2015, 7:18pm

suman · January 5, 2016, 12:23am

Happy New Year everyone! Let’s start it off with 3.8-10.12 update.

There are two main points to mention about this update. We’ve added Discourse(the software that runs this forum) as a Rock-On. Some of you may not know it, but this forum runs as a Rock-On and now you can too! I plan to write more on this topic separately. Secondly, our friendly moderator @phillxnet has patiently took input from users in the community and improved S.M.A.R.T reporting.

github.com/rockstor/rockstor-core

no SMART info on MSA70

opened 03:12PM - 03 Nov 15 UTC

closed 10:59PM - 04 Jan 16 UTC

phillxnet

enhancement

Thanks to @titus_36 on the forum for reporting this. It seems there is no WebUI …SMART info displayed yet it is available on the command line. Reason seems to be that the info is differently formatted to what the current parser expects and there is much less of it. Original reporter is hopefully to send in no-smart-msa70-report.tar.gz file resulting form commands requested on the following forum thread:- http://forum.rockstor.com/t/no-smart-support-for-msa70/641 Please update the forum post on progress with this issue.

suman · January 5, 2016, 12:23am

suman · January 10, 2016, 1:27am

3.8-10.15 is now available. This is a big one, for a testing update. @ganti_priya submitted a large pull request revamping our dinosaur-era underscore templates with handlebars.js. For developers, this translates to more readable and maintainable code. For users, it means a bit more efficient and robust UI. I am pretty thrilled we got this done finally as it’s been on the list forever and the tech debt kept growing.

Additionally, we managed to improve Rock-ons some more. Plus, I’ve added JenkinsCI. Along with Discourse from the last update, I am happy to share such Rock-ons that we use in production here!

We are pretty close to 3.8-11 now. We are mostly testing now. I’d say no more than 1-2 testing updates.

suman · January 10, 2016, 1:29am

KarstenV · January 10, 2016, 12:40pm

I am seing a very nice speedup in the user interface. A definite improvment.

Nice job.

KarstenV · January 10, 2016, 2:07pm

I do think I found an error.

Again it has to do with the Plex Rock-on.

First I must say: Plex is running, I can stream files from it. It sees my media share.

But I do get this:

Where the interface does not show the share.

When I click to add a share (Add storage), I get this:

Notice the code that is visible.

If I click storage (to add a share) I get this:

I am actually unable to add a share to my plex server.

Good thing that the old one is still working

suman · January 10, 2016, 2:10pm

Thanks @KarstenV, looks like we missed a template there. Should be cleaned up shortly.

suman · January 11, 2016, 11:58pm

Another day, another testing update… 3.8-10.16 is now available. As we get closer to the stable release, the changes in this update are mainly regression fixes and hardening, cleanup stuff in general. A good and quick update, nevertheless. Here’s the log

github.com/rockstor/rockstor-core

minor regression in fake serial ui logic

opened 04:49PM - 10 Jan 16 UTC

closed 07:24PM - 10 Jan 16 UTC

phillxnet

Prior to 3.8-10.15 we tested for our internal fake-serial drive flag and indicat…ed it to be unreliable as we also do with null, empty, or serial == diskName however as of merge (pr) #1045 we also have to test for serial.length == 48 to catch our new flag mechanism of a 48 char fake-serial-uuid-derived-char-string. See commit in above referenced pr:- https://github.com/phillxnet/rockstor-core/commit/ce733be859141d32954b2af88f543b8ec601fe9a From an initial look this seems like a minor change to line 208 in file :- /src/rockstor/storageadmin/static/storageadmin/js/views/disks.js is required as the display logic is now handled by a handlebar helper "display_disks_tbody".

suman · January 11, 2016, 11:59pm

Should be fixed in 3.8-10.16 @KarstenV. Give it another try!

KarstenV · January 12, 2016, 7:32am

Its not completely fixed.

I don’t see the code any longer, and I am able to ad shares / map them.

But the share I allready mapped is not visible in the UI.

The share is accesible though:

And works in plex as well (Plex is setup for my native language):

So there is still a bug somewhere in the plex settings panel.

suman · January 13, 2016, 3:34am

We have another testing update, thanks to contributions from @phillxnet and @mchakravartula. 3.8-10.17 is now available.

Important fixes and improvements to NUT, S.M.A.R.T and Access key management. Best part about this update? I did not write a single line of code

github.com/rockstor/rockstor-core

NUT communication error: configuration unreadable

opened 10:53PM - 29 Dec 15 UTC

closed 02:52AM - 13 Jan 16 UTC

peterdenhaan

On a fresh install of Rockstor (3.8-9.02) I had problems getting UPS support to …work until I fixed group ownership in `/etc/ups/`. I cannot find an issue for this here in Git, so I'm opening one. I am using an USB connected APC Smart-UPS 750 with the `usbhid-ups` driver. Looking through the journal, `upsdrvctl` started the driver just fine, but `upsd` was not initialising properly: ``` Dec 29 22:22:59 animalfarm systemd[1]: Starting Network UPS Tools - power device driver controller... Dec 29 22:22:59 animalfarm systemd-tmpfiles[5344]: Failed to open '/etc/tmpfiles.d/nut-run.conf', ignoring: No such file or directory Dec 29 22:22:59 animalfarm upsdrvctl[5347]: Duplicate driver instance detected! Terminating other driver! Dec 29 22:22:59 animalfarm usbhid-ups[5092]: Signal 15: exiting Dec 29 22:23:05 animalfarm upsdrvctl[5347]: interrupt pipe disabled (add 'pollonly' flag to 'ups.conf' to get rid of this message) Dec 29 22:23:05 animalfarm upsdrvctl[5347]: Using subdriver: APC HID 0.95 Dec 29 22:23:05 animalfarm upsdrvctl[5347]: Network UPS Tools - Generic HID driver 0.38 (2.7.2) Dec 29 22:23:05 animalfarm upsdrvctl[5347]: USB communication driver 0.32 Dec 29 22:23:05 animalfarm upsdrvctl[5347]: Network UPS Tools - UPS driver controller 2.7.2 Dec 29 22:23:05 animalfarm systemd[1]: Started Network UPS Tools - power device driver controller. Dec 29 22:23:05 animalfarm systemd[1]: Starting Network UPS Tools - power devices information server... Dec 29 22:23:05 animalfarm systemd[1]: Started Network UPS Tools - power devices information server. Dec 29 22:23:05 animalfarm usbhid-ups[5371]: Startup successful Dec 29 22:23:05 animalfarm upsd[5372]: fopen /var/run/nut/upsd.pid: No such file or directory Dec 29 22:23:05 animalfarm upsd[5372]: listening on 0.0.0.0 port 3493 Dec 29 22:23:05 animalfarm upsd[5372]: listening on 0.0.0.0 port 3493 Dec 29 22:23:05 animalfarm upsd[5372]: Can't open /etc/ups/ups.conf: Can't open /etc/ups/ups.conf: Permission denied Dec 29 22:23:05 animalfarm upsd[5372]: Can't open /etc/ups/ups.conf: Can't open /etc/ups/ups.conf: Permission denied Dec 29 22:23:05 animalfarm upsd[5372]: Network UPS Tools upsd 2.7.2 ``` The journal subsequently filled up with connection failure messages every 15 seconds (retries): ``` Dec 29 22:26:03 animalfarm upsmon[6067]: UPS: ups@localhost (master) (power value 1) Dec 29 22:26:03 animalfarm upsmon[6067]: UPS [ups@localhost]: connect failed: Connection failure: Connection refused Dec 29 22:26:03 animalfarm upsmon[6067]: Communications with UPS ups@localhost lost ``` The clue is in the first journal section above: `/etc/ups/ups.conf` cannot be opened by `upsd`. That looks like a problem. I tried to figure out why: ``` -rw-r----- 1 root root 1616 Dec 29 22:19 nut.conf -rw-r----- 1 root root 3784 Dec 29 22:19 ups.conf -rw-r----- 1 root root 4655 Dec 29 22:19 upsd.conf -rw-r----- 1 root root 2235 Dec 29 22:19 upsd.users -rw-r----- 1 root root 15708 Dec 29 22:19 upsmon.conf -rw-r-----. 1 root nut 4631 Dec 29 22:19 upssched.conf ``` Unless `upsd` is run as root - I hope not - it will indeed be unable to read its configuration file. On a hunch, I changed group ownership of all the files above to `nut`. The UPS code now is happy as far as I can see: ``` Dec 29 22:27:08 animalfarm upsdrvctl[6315]: Network UPS Tools - UPS driver controller 2.7.2 Dec 29 22:27:08 animalfarm systemd[1]: Started Network UPS Tools - power device driver controller. Dec 29 22:27:08 animalfarm systemd[1]: Starting Network UPS Tools - power devices information server... Dec 29 22:27:08 animalfarm systemd[1]: Started Network UPS Tools - power devices information server. Dec 29 22:27:08 animalfarm usbhid-ups[6317]: Startup successful Dec 29 22:27:08 animalfarm upsd[6318]: listening on 0.0.0.0 port 3493 Dec 29 22:27:08 animalfarm upsd[6318]: Connected to UPS [ups]: usbhid-ups-ups Dec 29 22:27:08 animalfarm upsd[6318]: fopen /var/run/nut/upsd.pid: No such file or directory Dec 29 22:27:08 animalfarm upsd[6318]: listening on 0.0.0.0 port 3493 Dec 29 22:27:08 animalfarm upsd[6318]: Connected to UPS [ups]: usbhid-ups-ups Dec 29 22:27:08 animalfarm systemd[1]: Reloading. Dec 29 22:27:08 animalfarm systemd[1]: Starting Network UPS Tools - power device monitor and shutdown controller... Dec 29 22:27:08 animalfarm systemd-tmpfiles[6329]: Failed to open '/etc/tmpfiles.d/nut-run.conf', ignoring: No such file or directory Dec 29 22:27:08 animalfarm systemd[1]: Started Network UPS Tools - power device monitor and shutdown controller. Dec 29 22:27:08 animalfarm upsmon[6331]: fopen /var/run/nut/upsmon.pid: No such file or directory Dec 29 22:27:08 animalfarm upsmon[6331]: Using power down flag file /etc/ups/killpower Dec 29 22:27:08 animalfarm upsmon[6331]: UPS: ups@localhost (master) (power value 1) Dec 29 22:27:08 animalfarm upsd[6318]: User monuser@127.0.0.1 logged into UPS [ups] Dec 29 22:27:08 animalfarm upsd[6318]: User monuser@127.0.0.1 logged into UPS [ups] ``` Given that this was a clean Rockstor install, I suspect that group ownership is not correctly set when these configuration files are first installed or generated. Thanks for reading, and thanks for the amazing work you guys are doing on Rockstor.

suman · January 13, 2016, 3:36am

suman · January 21, 2016, 9:57pm

I don’t understand this part. What’s the name of the Share? Shares are not mounted under /media. So, if you are referring to /media/plex being the Share, it is not correct. Or am I missing something here?

suman · January 21, 2016, 11:52pm

3.8-10.18 is now available. This may very well be the last Testing update in this cycle. We can expect the Stable release as quickly as tomorrow.

I am wrapping up improvements to Rock-Ons for this cycle with this update. There are some UI related changes that I wished to have included in this cycle, but due to UI template overhaul, I’ve decided to tackle them in the .12 cycle. Thanks to @phillxnet 's contributions, a couple more issues got resolved. The most exciting news of this update, however, is that we have a new contributor! Thanks Mazo!

github.com/rockstor/rockstor-core

Share Access control issue found on latest unstable/test updates

opened 01:37AM - 14 Jan 16 UTC

closed 01:55AM - 17 Jan 16 UTC

parker1c

bug

I believe I found some bug with "Access control" in a share a user makes and the…n edits. I can repeat the issue as I wasn't sure what broke until after some testing. Symptoms: When editing a created share's permissions,(in the webgui) for example: enabling read/write/execute for all 3, owner/group/other (for example 777). Upon saving, it then goes to read only for owner/group/other (444), despite selecting other options. Default install of 3.8-10(Linux: 4.2.5-1.el7.elrepo.x86_64) ISO does not exhibit this behaviour. Upon activating Testing updates, and installing/rebooting (Rockstor 3.8-10.17) (Linux: 4.3.3-1.el7.elrepo.x86_64), the above symptoms now arrive. Installing updates after clean install are only things done. New or old shares created and then edited now present the "read only" option despite selecting other options for access control.

github.com/rockstor/rockstor-core

gui smart test unavailable when last test aborted by host

opened 04:30PM - 13 Jan 16 UTC

closed 02:11AM - 16 Jan 16 UTC

phillxnet

If the last smart self test was aborted by the host the "Perform Tests" UI logic… interprets this as a test in progress. Consequently it does not offer any test options and leave the user unable to perform further tests. A message "The self-test routine was aborted by the host." is displayed but this is redundant as it is also available in the Capabilities tab and the Self-Test Logs tab. I am having a quick look at this now.

suman · January 21, 2016, 11:57pm

KarstenV · January 22, 2016, 6:42am

The share is my media share on my btrfs pool.

To make it accesible I used the Add Storage function available in settings for the plex rock-on (I have put in the values I choose, as an example).

This, I believe, maps a share to /media/plex on my root drive pointing toward the share on my BTRFS pool.

Earlier this share mapping was visible from within the plex settings, but it isn’t anymore.

Hope this clears things up.

suman · January 23, 2016, 12:01am

3.8-11 is now released! Announcing Rockstor 3.8-11