Printable View
So just to double check.....this ONLY affected Untangle rigs that were on some prior 15.1.0 version....if they punched in an update and rebooted, it would blow up. Or say..they went and rebooted?
It would NOT affect Untangle units currently at 15.0.0? Or say...14.2.2? Those would be safe to now update and reboot?
Just want to find some way to know which rigs might be safe to reboot....got a lot out there in production.
If you are on 15.0 or 14.2 they should be fine, this was only affecting a 15.1.0 build. Though they may not have the update to 15.1.1 available without adding to our early access list. If you get to 15.0 and its not allowing you to update further send in a support ticket with the UID you need pushed and we can add those.Quote:
Originally Posted by YeOldeStonecat
Thanks Shawn (It's Brian..I had a ticket with you earlier this week re: something else..."upgrade process interrupted". Although..may be related to this issue).
Had 4x rigs blow up ..fail to reboot...in the past 7-10 days.....assumed tanked hard drive, but now I guess was this one.
I can only answer half of your question:Quote:
Originally Posted by donhwyo
Attachment 10592
...which isn't an answer, at all.
If you uncheck 'Connect to Command Center', all cloud management for the box would be disabled.Quote:
Originally Posted by donhwyo
I personally would advise against that. Even in this unfortunate case, where VPN connections were dropped for a few seconds due to forced upgrade, the alternative is that your box could be bricked. This is in addition to not having any Command Center functionality to manage the box (remote access, reports, policy management, etc).
I suppose what's necessary is necessary, and what's done is done...Quote:
Originally Posted by Timur
perhaps it would've been better to do the push a few hours later? 02:00 UTC would've been off-hours for the entire western world, which is presumably where the vast majority of your boxes live. That would've been far less noticed than the middle of the day.
When the push happens isn't instant, there's a delay. I had units upgrade at midnight as per their usual schedule here, and I had others update in the middle of the day TODAY, two days later. Because the sad part is pushes are almost never pushes, just really aggressive pulls. But I digress...
If you've ever worked with RMM tools you'll be familiar with this sort of lag, it simply exists. And despite all efforts, in many ways defies scheduling. As it was Untangle pushed this update so hard I was having intermittent issues connecting to updates.untangle.com for the last 24 hours.
One does not simply push a patch to thousands of units within the 2 hour window that's generally easy for the US, but is also not a train wreck for Australia.
Microsoft does this to us at least twice a year! The fact that it happened to me is irrelevant, I'm only upset there wasn't an announcement made BEFORE the event. I should have been able to come here and find someone making a post about it, or been making the post myself because I saw it in the change log. The event is irrelevant, these things happen. The issue is the lack of transparency before the event was created.
I'm still not sure specifically which 15.1.0 date release contained the bug.
It is easier to ask forgiveness than permision. LOL
It's also easier to keep an existing customer, than to lose one and have to gain a replacement one.Quote:
Originally Posted by donhwyo
Said repeatedly in the business world.
I reloaded two Z4w boxes myself. We're lucky we only had two, and we have since turned off auto-update on all of our Untangle boxes (we are a partner, and manage 50-80 of them). Those two were the same client, which wassn't particularly helpful as they were separate branch offices at a distance, running a BOVPN (ouch).
Compounding this with updates we don't approve being pushed is quite serious. I'm on Untangle's mailing list. If I'd had an e-mail that says "We have no choice, but need to....because...we will do it at this time... in an e-mail would have been something I'd have expected from every other SMB firewall vendor I've worked with. Sonicwall. Watchguard. And so on.
We're at water over the dam point now, but this really isn't a good look for a company that goes beyond just individuals and extends into the SMB market. There needs to be a redesign of how Untangle communicates with its customers, not just the updating process itself, to prevent this from ever happening again, because without that redesign, if things happened this way again, I would ensure I moved all of our clients to another solution.
Hopefully they learned from the mistake. Since 16.0 is an RC we can only hope.
I have only had one bad upgrade since version 5. It was very long ago. As usual it was at about as bad a time as possible. I have turned off auto upgrade on everything ever since. I also have a spare box ready to switch over to if needed.
Just because you know you are paranoid doesn't mean they are not after you.
