Set V2 to Record to U4 HDD

[Mork]

My last DP Series (DP-Lite) is getting unstable, I am sure it is probably just old electrolytics, but I thought I would upgrade to a V2, so have just ordered one (Triple Tuner, no HDD). My attention now turns to the V2. (I have a U4 in our main room)

I have a spare 500GB SDD that I could task to the V2, but I have seen a couple of posts that mention that recordings can go to a U4. So my thinking was to use a spare 32GB SD-Card for TimeShift, and send all recordings to the U4. (I have a wired GigE Network)

So some questions:
1. Is that possible?
2. Does it work reliably?
3. Any issues playing recordings made on the U4?

Looking forward to my new toy coming next week!

[MrQuade]

No issues with that setup at all. It's a great way to go, and both V2 and U4 will be able to play both sets of recordings.

As long as the U4 is always on our in standby, v or otherwise available on the network, the V2 will happily use it as storage.

Just try not to record three same thing on both boxes to the same directory at once. You could end up with both of them trying to create the same file at once. Not sure if the firmware can account for and compensate for that situation.

[Paul_oz53]

No issues with that setup at all. It's a great way to go, and both V2 and U4 will be able to play both sets of recordings.

...
Just try not to record the same thing on both boxes to the same directory at once. You could end up with both of them trying to create the same file at once. Not sure if the firmware can account for and compensate for that situation.

To avoid this problem you could set a greater prepadding on one machine (assuming you use standard or long filenames) to ensure that the filenames are actually different.

[Grumpy_Geoff]

No issues with that setup at all. It's a great way to go, and both V2 and U4 will be able to play both sets of recordings.

...
Just try not to record the same thing on both boxes to the same directory at once. You could end up with both of them trying to create the same file at once. Not sure if the firmware can account for and compensate for that situation.

To avoid this problem you could set a greater prepadding on one machine (assuming you use standard or long filenames) to ensure that the filenames are actually different.

Or perhaps set the default recording location on the V2 to be a "V2" sub-directory on the U4's disk.

[prl]

Just try not to record the same thing on both boxes to the same directory at once. You could end up with both of them trying to create the same file at once. Not sure if the firmware can account for and compensate for that situation.

To avoid this problem you could set a greater prepadding on one machine (assuming you use standard or long filenames) to ensure that the filenames are actually different.

The system should make sure that the filenames are different even if they have the same program name, channel name and timestamp part (this is essential if the user chooses "short" recording name format - that only has a date timestamp, and no part of the filename contains the time of day).

The filename generation should not have any race conditions, because this part of recording creation is sequential, even for timers that start at the same time.

For those interested, the recording system puts a _NN just before the filename extensions for files after the first one with that base name: i.e. _00 is never added. It's handled in getRecordingFilename() in Tools/Directories.py.

[Paul_oz53]

Just try not to record the same thing on both boxes to the same directory at once. You could end up with both of them trying to create the same file at once. Not sure if the firmware can account for and compensate for that situation.

To avoid this problem you could set a greater prepadding on one machine (assuming you use standard or long filenames) to ensure that the filenames are actually different.

The system should make sure that the filenames are different even if they have the same program name, channel name and timestamp part (this is essential if the user chooses "short" recording name format - that only has a date timestamp, and no part of the filename contains the time of day).

The filename generation should not have any race conditions, because this part of recording creation is sequential, even for timers that start at the same time.

For those interested, the recording system puts a _NN just before the filename extensions for files after the first one with that base name: i.e. _00 is never added. It's handled in getRecordingFilename() in Tools/Directories.py.

I understand this to be true if one box is recording the same thing twice but will it be true if two boxes record to the same directory?

Of course, if there is not a critical race condition then the laggard will detect the existing file so I imagine the risk of collision may be too small to be a problem in normal use.

[Grumpy_Geoff]

The system should make sure that the filenames are different even if they have the same program name, channel name and timestamp part (this is essential if the user chooses "short" recording name format - that only has a date timestamp, and no part of the filename contains the time of day).

The filename generation should not have any race conditions, because this part of recording creation is sequential, even for timers that start at the same time.

But the timers start on different boxes, so is it still "race safe"?

For those interested, the recording system puts a _NN just before the filename extensions for files after the first one with that base name: i.e. _00 is never added. It's handled in getRecordingFilename() in Tools/Directories.py.

It's 3 digits in my experience, not that we're pedantic at all on this forum

[Mork]

Thanks everyone. It's great to be back in this forum and see so much help being offered.

I was going to set up the pre padding differently as a potential mitigation for a race condition anyway.

When I get the box and set all this up I will test out a few use cases and report back with any problems.

Although technically this will be a 7 tuner setup. (From the perspective of the U4's HDD) I wont try to see how 7 recordings and 1 time shift stress the HDD, there is just never that much worthwhile recording on FTA now anyway.

[Paul_oz53]

...

Although technically this will be a 7 tuner setup. (From the perspective of the U4's HDD) I wont try to see how 7 recordings and 1 time shift stress the HDD, there is just never that much worthwhile recording on FTA now anyway.

I predict this will not be a problem - I have recorded 12 services to my U4 with no problems so 7+1 shouldn't be an issue.

[MrQuade]

One other thing you can do is set up the"other" wiz as a fallback tuner.

That way, if either one runs out of tuners, it can use the network connection to the other wiz and borrow one of its tuners.

[prl]

The system should make sure that the filenames are different even if they have the same program name, channel name and timestamp part (this is essential if the user chooses "short" recording name format - that only has a date timestamp, and no part of the filename contains the time of day).

The filename generation should not have any race conditions, because this part of recording creation is sequential, even for timers that start at the same time.

But the timers start on different boxes, so is it still "race safe"?

No, it's not. Brain not engaged before fingers applied to keyboard.

Tweaking the pre-padding will help with this, but if AutoTimers are used, they may adjust the start times so that if a program that is scheduled to start at 8:00 actually starts at 8:01, then that may interfere with another timer with the same scheduled start time and one-minute different padding and actually starting at 8:00. IceTV timers should work OK wit a small tweak to the padding, because the IceTV EPG (and timers) are always rounded to an even multiple of 5 minutes.