Rockstor.com maintenance

Sorry folks but I’m having to do a little server maintenance so will be rebooting the server hosting rockstor.com.

The consequence of the inevitable downtime is expected to be a couple of hopefully short outages (few minutes each) for the main web site, Docs included, and for the Rock-ons repository. We also host a very old CentOS only variant of the “Rockstor-Base” rpm repository there. But this having a short downtime is not thought to present an issue to existing old installs.

The EULA linked to in the installer setup page is also hosted on this server here:
http://rockstor.com/eula.html
which currently reads as follows:

Rockstor EULA

Rockstor comes with no guarantees or warranties of any sorts, either written or implied.

The Distribution is released as GPLv2. Individual packages in the distribution come with their own licences. A copy of the GPLv2 license is included with the distribution media.

It’s unavailability should also not present a problem other than the link being broken for the server reboot periods (copied in above so we don’t ask folks to tick for an uncertain reason). And one last element is a redirect from our old forum link of:
http://rockstor.com/forums (which like the old repo is only used in very old installs within the Web-UI.)

This forum is otherwise hosted on an independent server, as is Appman and our Subscriptions system. But building from source will also be affected due to a couple of jslib zips retrieved from rockstor.com during a source build: this single point of failure in source builds should ideally be fixed; all in good time however.

So in short expect a little down time for:

- Main web page.
- Docs.
- Very old rpm repo in CentOS variants.
- EULA link in setup page.
- Source builds as lib.tgz will be unavailable.

The lib.tgz is simply a tar zipped copy of our rockstor-jslibs repo here:

I’ll post again here once I’ve completed this latest ‘phase’.

Thanks for you patience; we try to keep things simple on the back-end but that in turn comes with some ‘old’ woes.

3 Likes

That’s probably the first time I’ve actually read the EULA :slight_smile:

Shame on me :wink:

2 Likes

Linking to an outstanding related issue re moving to https for our main web site / rock-ons etc, currently in rockstor-core as we don’t yet have GitHub repo for our main web site: news on this soon hopefully however:

As always, if you notice something astray re this parallel https setup then feel free to comment here or on the linked GitHub issue. Best we run both in parallel for a bit to enable wider examination / testing before I do the final permanent redirect of http -> https.

Cheers.

3 Likes

OK so we now have

in place.

Tests of expected functions of the affected services:

Main web page

Visiting any of the now deprecated http pages now ends with a permanaent (301) redirect to the similarly located https. Confirmed in browser and on server logs, i.e. 301 followed by a 200.
N.B. we have some very old hard-coded http references within our page source, these don’t look critical and can be addressed once we open the proposed new GitHub repo to host our main web site to ease contributions/correction.

Docs

Same as for main web page, redirects appear functional, and most internal links are thought to be relative anyway given we use sphinx to construct our html from:

Rock-ons

wget http://rockstor.com/rockons/root.json

responded as expected with indicated 80 -> 443 redirect and re-request and consequent successful download.
A prior CentOS based Rocksto 3.9.2-57 was also able to successfully update it’s rock-ons using the new server config.

N.B. we currently fetch root.json and then every json mentioned within it. And as these now incure a re-direct/re-connect we should probably look to changing the rockstor-core code on this front. Hopefully these’s initial permanent redirect 301’s arn’t going to be too costly on our server. We will have to wait and see on that front.

On a side note we might want to enable server side compression (with caching hopefully) to gain back some of the bandwidth/load we may have lost here. But this is material for another forum post/GitHub issue.

Old Legacy CentOS only repo

Defined in /etc/yum.repos.d/Rockstor-Base.repo on Rockstor 3 CentOS variants.
repo id = rockstor
repo name = Rockstor 3 - x86_64
Tested via:

yum clean expire-cache
yum --disablerepo="*" --enablerepo="rockstor" update
rockstor                 | 2.9 kB  00:00:00

Web server logs confirm a 301 followed by a 200 from yum.

EULA link

As per docs this is redirected to the new https variant.

JSLIBS lib.tgz

https://github.com/rockstor/rockstor-core/blob/master/buildout.cfg#L181
i.e.:
url = http://rockstor.com/downloads/jslibs-dev/lib.tgz

wget http://rockstor.com/downloads/jslibs-dev/lib.tgz

As per our Rock-ons test this shows the expected 301 redirect and we finally get our expected file.

So I’m calling this one done for now. Please comment here, or in the above linked GitHub issue if you find something I’ve missed.

I’ve got a little more tidy-up to do on this server but it should now be more service as usual hopefully.

Thanks for the understanding re a couple of 2 mins max outages incurred on this one.

3 Likes

That’s probably 4 minutes more downtime than my Rockstor installs have had :slight_smile:

And now i know the EULA.

Nice work @phillxnet

1 Like

Sorry about this but I’m going to have to down this server for around 20 minutes. All now looks to be working as-is so I want to ensure we can at least return to this point if something untoward were to visit us.

Scheduled down time 11:00 am UTC +1 15-20 mins.

Will update here once that’s done and hopefully we can then call this one ‘fixed’. At least until the next round of old school ‘maintenance’.

Thanks again for your patience and understanding.

OK, so we were down around 19 minutes. I’ll leave this server be for a bit now, systems wise. All prior services look to be OK.

Again apologies for these interruptions.

2 Likes

OK folks we are again at a maintenance juncture.
Apologies for any inconvenience as there will be downtime in the previously specified services.

Scheduled intermittent down time from 10:30 am UTC +1 for around 2.5 hours.

Thanks for your patience and understanding during these outages.

All done now, at least for the time being.

We were down, during the specified period, for around 45 mins in total. Mainly in two blocks of 20 mins.

All now looks to be back as normal again, only a little tiny bit shinier.
Again apologies for these interruptions in our services.

2 Likes