Nightly reboots: Why they must go...

[John Kotches]

Never implied that that was the case. Just saying, Hopper is the fastest sat receiver on the market. I agree, especially with *NIX boxes, reboots aren't mandatory by any means with good programming.

Sent from my iPad 2 using Forum Runner

Rather than replying twice I'll say that 95% isn't really that good. That's 3 minutes for every hour of operation down. Something ain't right, and they've had more than a year to fix it.

Add in the bugs that seem to appear with each new release and from my perspective they have a QA/QC issue. The new bug is a sudden disconnect of a Joey from its assigned hopper. As best I can tell its only happening with DVR / paused playback. I don't recall this happening "live".



Sent from my Samsung Galaxy Note 2 using Tapatalk 2.x

 

[3HaloODST]

Rather than replying twice I'll say that 95% isn't really that good. That's 3 minutes for every hour of operation down. Something ain't right, and they've had more than a year to fix it.

Add in the bugs that seem to appear with each new release and from my perspective they have a QA/QC issue. The new bug is a sudden disconnect of a Joey from its assigned hopper. As best I can tell its only happening with DVR / paused playback. I don't recall this happening "live".

Sent from my Samsung Galaxy Note 2 using Tapatalk 2.x

I agree that 95% is bad. However unplug your EHD before the nightly reboot to take the reboot time from over an hour to less than ten minutes. Record stuff (and delete later) starting at 1am to delay the reboot. Yes this needs to be fixed but refusing to do the workarounds then complaining about it taking so long isn't going to get it fixed any quicker.

Sent from my iPad 2 using Forum Runner

 

[KKlare]

Someone mentioned our $1000 921. It would be lucky to go one day without a reboot if you did any activity like looking at the guide paged over several days and many channels. It think the 722s and Hopper will go for 2 or 3 days. Yeah, memory leaks are a likely problem. I have had the Hopper on at least one occasion have problems responding when fsck was not done for a couple of days. BTW a reboot does not force an fsck (no way I know to do that within the Hopper) but the time of the fsck will be next to the reboot. The best seems to set to "off" the Hopper and all its Joeys after 1 AM or any time after that for more than 1/2 hour, best guess.

John, it sure does not seem like a journaled system with my limited knowledge of them. It does not permit fragmented files, which you claim for the journaled system, for example, but they surely work on the internal. The good thing is that you can break a transfer will no noticeable chance of losing the file from both, although I think that has happened to me with too little documentation to back up my claim.

Humbug, just had an hour program on the internal break into 17 and 43 minutes with no reboot, obviously or would have lost at least 5 minutes. How? Oops two other 1/2 hour broke at the same time into 17 and 13 minutes, all non-PTAT. Curiouser.
-Ken

 

[khemka]

John

Thanks for bringing this to my attention. Clearly the intention is for the process to be 10 minutes or less - it is not just a nightly reboot, but there are many other critical checks processes that are kicked of by the nightly reboot.

Let me see how we can improve it for users with EHD's and will get back to you on this forum here

Vivek

Some think I have been extremely (and overly) critical of the Hopper for the nightly reboots. I would argue that I set higher standards. I'm paying for a service and the devices aren't available a significant portion of the time. If my systems only had uptimes of 95.x% they wouldn't be my systems any longer; I'd be unemployed. Even standalone systems should be able to maintain 99% uptime. Clearly that isn't the case with hopper.

During the hopper reboot, no recordings can kick off and any attached joeys are unavailable. Yes, it's done in the middle of the night but I've missed recordings because the nightly reboot doesn't complete quickly enough. If you have 2 hoppers, your still SOL, as they both reboot at about the same time.

I have managed various flavors of Unix systems for 25 years. In those 25 years I haven't had any systems that required daily or even weekly reboots. I have encountered poorly written applications which required restarts on a scheduled basis due to memory leaks. But that didn't require a reboot of the OS to clean up, after all it's Unix of some form. I currently manage ~150 nodes with ~250 unique OS instances and in the vicinity of 350 TB of user data in a directly customer facing application.

Hopper is in a directly customer facing application.

Vivek:

If you have any background in IT and specifically in service delivery you should be appalled by this. At this point I want an explanation -- why must these be rebooted nightly and why does a system check that you obviously think should take 10 minutes (or you wouldn't put it up there) taking 60+ minutes. Since the hopper is a purpose built computer running some flavor of a *nix OS on a MIPS processor I expect it to behave like one.

Your product is in a directly customer facing role, and you need to treat it as though it is. Currently from my perspective you are not.

 

[3HaloODST]

John

Thanks for bringing this to my attention. Clearly the intention is for the process to be 10 minutes or less - it is not just a nightly reboot, but there are many other critical checks processes that are kicked of by the nightly reboot.

Let me see how we can improve it for users with EHD's and will get back to you on this forum here

Vivek

Please, even if nothing else gets done about this, at the very least, can we have the option to choose the time for the nightly reboot?

 

[khemka]

Yes - that is being actively worked on. I cannot commit to a timeline - but it is being worked on. I will have updates on both

Vivek

Please, even if nothing else gets done about this, at the very least, can we have the option to choose the time for the nightly reboot?

 

[Scott Greczkowski]

How's that for service guys?

Thanks Vivek!

 

[navychop]

Good news.

 

[charlesrshell]

How's that for service guys?

Thanks Vivek!

Nice

 

[dare2be]

Don't you wish you had called Dish directly now?

 

[dwarren2]

I hope Vivek's team can fix the reboot time. I've lost some scheduled recordings on the Hopper with the 2tb ehd because it takes so long to reboot. I'm not usually up at that time, but the 1 time I was, it took over 1.5 hours even though it kept saying it should only take 15 minutes. I've learned that if I want to record something in the early morning, I use the other Hopper that has a 750 gb ehd.

 

[KE4EST]

How's that for service guys?

Thanks Vivek!

This is why I love Satelliteguys!

 

[dwarren2]

Vivek, if you are still folowing this thread, here's another request. with my 722, I could access my ehd remotely through my slang adapter. That went away with my Hopper. DIRT said that that was a bug in the 722. Please bring back that bug.

 

[John Kotches]

John

Thanks for bringing this to my attention. Clearly the intention is for the process to be 10 minutes or less - it is not just a nightly reboot, but there are many other critical checks processes that are kicked of by the nightly reboot.

Let me see how we can improve it for users with EHD's and will get back to you on this forum here

Vivek,

I appreciate your taking the time to reply, but you're going to have a tough sell. Keep in mind when I type you or your I mean Dish not you personally.

If you are not upgrading the firmware, what is so broken that your solution is to reboot? Hopper is a server which provides streaming services to one or more streaming clients. The minimum it's serving is itself. That means not rebooting it as a fix for sloppy code as a fix all.

The reason it has to run these checks is that you called for a reboot. The system dutifully follows the instruction to shutdown. If you aren't upgrading firmware, which should be child's play to determine, there shouldn't be a need to reboot.

I'd happily discuss this off line of you would like to as well.




Sent from my Samsung Galaxy Note 2 using Tapatalk 2.x

 

[CaroleV]

Sorry, but I have been hearing from Dish that this reboot thing is being "actively worked on" for almost a year now and nothing has been done. As a result, I am "actively working on" getting through my two year contract so I can get rid of this nightly annoyance. Even though I have the thing set to record from 1 - 2 AM, the Joey in my bedroom still reboots between 1 and 1:30 and I have not been able to find a work-around. It would not even be quite so irritating if the show you were watching came back on after the reboot, but noooo can't even have that...the stupid screen saver comes on! It's insane.

 

[John Kotches]

Carole,

Between this and the Joey disconnects my wife is getting dissatisfied with Dish.

Sent from my Samsung Galaxy Note 2 using Tapatalk 2.x

 

[MikeD-C05]

All DISH receivers reboot around 1:00am to help clear out any problems , do housekeeping , download the newest guide, etc. WIthout , you would suffer more random reboots during the day time, when you are actually awake to watch tv. I've had them for over 16 &1/2 years and I don't even think about it anymore.

 

[Bobby]

The big difference Mike is that, with other receivers, you have the option to change the reboot time. Not so with the Hopper.... Something I've thought about, but have not seen mentioned, with regard to the 1AM reboot is that Auto-Hop begins after 1AM the following day for PTAT. I wonder if this has anything to do with why it is necessary to have a daily reboot.....

 

[John Kotches]

Mike,

We will always disagree that a daily reboot is necessary for housekeeping.

Hopper is a server for streaming services to one or more clients. Servers shouldn't require daily reboots and fsck.

Perhaps I should stop thinking about this as Dish providing a service that has an expectation of greater than 95% availability. A service is only as good as the lowest availability in the service delivery chain after all.

How many of us would pay for services where the availability was 95%?



Sent from my Samsung Galaxy Note 2 using Tapatalk 2.x

 

[dare2be]

How many of us would pay for services where the availability was 95%?

Even paid escorts have to sleep sometime. :