Hi there,
My NS8 instance failed to backup for some days (storage down) but I didn’t received an email notification.
Is this expected ? Is there a preferred way to actively monitor a NS8 server for critical issues like this one ?
Thanks for helping.
Matthieu
Hi Matthieu,
Currently, NS8 itself does not send email notifications for backup failures.
However, monitoring, alarms, and email notifications are available through Nethesis cloud services for clusters with an active subscription.
If you’d like to implement a custom notification script, you can use the APIs to list failed backups. For example:
api-cli run list-backups | jq -r '.backups[]|.instances[]|select(.status.success==false)|.module_id'
The APIs are especially useful if you want to test from a remote host or gather detailed information.
For a simple success/fail check on the NS8 leader node, you can inspect the contents of this file:
cat /run/backup-monitor.dat
An F in the output indicates that one or more backups have failed.
Feel free to reach out if you have any questions or need further help!
Hi
Are there any plans to support NS8 in Debian? I get this message in the subscription section.
Unfortunately not at the moment. Read here.
It IS possible to set up a Rocky Instance, and migrate all Apps from Debian to Rocky, which IS supported.
At least a “backdoor” for a subscription…
My 2 cents
Andy
Hi Andy,
Why make it complicated when it can be easy? With a corresponding subscription for Debian, for example. In my opinion, one of the most stable Linux OS should not be left out in the long run.
Have a nice Sunday…
Uwe
Hi @transocean
Don’t need to ask me, I’ve been Team Debian long before Rocky was dreamt about…
Have a great Sunday.
My 2 cents
Andy
Thanks Andy; I’m actually more comfortable with Debian; sure I can learn to use Rocky but switching my way of working between distributions is a bit tiring. Upgrading or installing NS8 is already a complex task at the beginning.
I’m going to keep using Debian for what I’m practicing and as soon as I feel I understand the basics I’ll try Rocky.
I hope and this backdoor will remain by then in case I need it.
Regards
All my customers are using Debian, I’ll probably keep things this way. I do not want Rocky, I do not need clones of RedHat bugs.
Their claim to fame is:
A bug for bug clone of RedHat Enterprise Linux.
My 2 cents
Andy
Because our support team has a strong know how on RHEL based system.
Right now, we are note able to give the same support level on Debian systems.
I understand that. But in the long term, you should be more diversified.
That’s your call, of course, but it’s obvious that Debian is a second-class citizen in terms of NS8 support, and I suspect it’s unlikely that’s going to change any time soon (if ever).
But to the subject of this thread, NS8’s error handling is inexcusably poor, and the backup situation is a good example of it. Not only does it not email the admin on a backup failure (which it ought to do, without a subscription or any kind of custom scripting–and it doesn’t do even with a subscription), it lies to you on the cluster status page and tells you everything is fine with the backup. All it means is that you’ve configured one, not that it actually works. To find out if there have been errors with the backup, you have to browse to the backup page, find the scheduled backup you’re interested in, and click on See Details–then you’ll see if the last backup succeeded. A handful of apps will report this on their own page, but there’s no consistency there.
@davidep, this is just unacceptable. If a backup fails, NS8 needs to notify the admin directly. It needs to show this clearly on the main cluster status screen, and if the system’s capable of sending email, it needs to do so. With as much as it complains about backups not being configured for any app, there’s just no excuse for this shortcoming. NS7 was far from perfect in its error handling, but at least it got this right.
I tend to see things the other way around…
It seems that Rocky still pushes “designed to be 100% bug-for-bug compatible with Red Hat Enterprise Linux” as their major “Raison d’être” (Reason to be / exist).
As it’s per definition a 100% bug compatible clone of RHEL, I really wonder what is under intense development by what community?
→ Maybe they contribute to the packaging graphics? 
Debian, on the other hand, is still pushing the borders what a rock solid, stable Linux can achieve.
And it’s been doing that longer than Rocky, or even RedHat has!
In that sense, its obvious to me that Rocky Linux is a second-class citizen in terms of Linux and Open Source. It’s not enhancing anything, it priorises getting bugs cloned 1:1 correctly!
And if they keep their word (Not that I trust anything from RedHat, nor any clones), this will NEVER change - Rocky wants to remain a bug compatible clone!
I do however fully agree about the catastrophic error handling NS8 has, especially your exemplary Backup.
I can handle anything Debian throws at me - I can’t do that for everything NS8 throws at me, agreed. But I can and will solve the issue with Debian. If a Debian based NS8 issue can’t be solved with NS8 community I certainly won’t switch to Rocky…
As said:
I can and will solve the issue with Debian. If NS8 doesn’t want to join the journey, fine.
My 2 cents
Andy
I’m not in the least interested in arguing whether Debian or Rocky is the better distro. The fact is that Nethesis support Rocky, don’t support Debian, and IMO that’s unlikely to change. I’d be happy to be wrong here, but that’s how I see it at the moment.
Fact is: It doesn’t, no matter which OS is on top or “supported”.
…and that really is the problem. If backup is important–and the devs know it is, and NS8 obnoxiously warns the admin if there’s a single app for which it isn’t set up–then it’s no less important that the admin be notified if the backup fails. NS8 is perfectly capable of sending email regardless of the underlying OS.
That’s what any sensible admin would think it means, yes. But what it actually means is that you have a backup configured for all installed apps. Agree that it’s very misleading, as I noted above.
Actually, showing this “badge” on the page “Cluster Status” with “All good” written with a green tick implies the last backup worked, not the backup is correctly set up…
It’s worse than both of actually think at the moment!
This is VERY misleading!
…and as I think more about it, NS8 supports email notifications, and complains if you haven’t configured them. But I don’t think I’ve ever seen it send any…
Obviously I support that.
Thank you for reporting the issue with the Cluster Status page. It appears to be a bug: Failed backup not reported by Cluster Status · Issue #7241 · NethServer/dev · GitHub
Regarding backup and email notification integration, I agree that it is essential. I hope the Core will introduce a command like sendmail to simplify the implementation of email notifications across all NS8 applications.
On the topic of Backup and Subscription, the failed backup alarm should have been functioning correctly since Core 3.2.4 for Business and Partner/Enterprise plans: Backup alert for systems with Subscription · Issue #7173 · NethServer/dev · GitHub
