View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0012411 | MMW v4 | Synchronization | public | 2014-12-04 02:02 | 2014-12-12 23:31 |
| Reporter | peke | Assigned To | |||
| Priority | urgent | Severity | major | Reproducibility | always |
| Status | closed | Resolution | fixed | ||
| Product Version | 4.1.6 | ||||
| Target Version | 4.1.6 | Fixed in Version | 4.1.6 | ||
| Summary | 0012411: Playlist analyze and changes update takes too long | ||||
| Description | When MMA sends updated playlist MMW is rechecking whole playlist making sync analyze step to go for more than hour just to update metadata instead of just do the sync up to device capacity or inform that sync is to large. eg.: 1. I have Playlist in MMW named "Test" and it is Auto playlist showing my entire library of tracks that I have never played (63752 tracks, 1.2TB) Which I selected for sync 2. Create Playlist "Test" in MMA and add track to it 3. Initiate Sync Based on MMW log it looks like MMW is trying to analyze What will be synched with "Test" on MMA and serves each track making sync stalled in Analyze for 45 Min with no moneing more than few percents of progress bar. This indicate that MMA "Test" playlist was not paired with one in MMW and MMW is assuming it is the same one where instead it should recognize that MMA "Test" was static and not one in MMW by just avoiding sync or create new sync playlist and sync tracks. | ||||
| Tags | No tags attached. | ||||
| Fixed in build | 1724 | ||||
|
|
Peke, did you upload the log somewhere? I haven't found it on our FTP, could you upload the debug log? |
|
|
I created new LOG file on FTP as old one didn't show it clearly where slowdown is. In this log you can see that: - First sync finished within minute - Second sync was initiated after I created "test" playlist on device and added two tracks in it. a) Update Metadata was stalled for few minutes even first sync already updated metadata b) Analyze was not finishing and I terminated sync eventually - after Third sync is initialized I almost couldnt terminate sync even I initiated terminate imediately after sync started |
|
|
OK, I am starting to understand the issue. At first we should mention that we are talking about WiFi sync between MMA <-> MMW. Based on the log the slowdown is caused by the fact that in the second sync all your 63752 tracks were added to the sync-list, i.e. probably the 'Test' auto-playlist you mentioned. I see two issues here: 1) on MMW: If I understand correctly the auto-playlist 'Test' was auto-added to the sync list without your intervention, is that right? Or did you add the auto-playlist manually to the sync-list between the first and the second sync? 2) Even if the sync-list is much larger than the device capacity, MMA tries to read whole the sync-list beforehand (it takes 30 minutes for MMA in this case). I guess that MMA should do this in batches, e.g. process and download 10 tracks and continue processing until the device capacity is full. This issue is also related to 0012138 |
|
|
I found that issue 1) can happen when the device time and the PC time are not synced. Fixed in build 4.1.6.1724 Added note to 0012138 for Marek re issue 2) |
|
|
2) Wouldn't be better that MMW do the hardcut as loading Playlist on MMA to analyze would take time where MMW can hardcut playlist within seconds by adding Size criteria which will further optimize sync speed and not affect sync content. Note: Loading playlist in MMW takes 4 seconds and like you explained in 2) "MMA tries to read whole the sync-list beforehand" takes 30 Min which I doubt would take if MMW cut that list to device free space size and than send it as sync list to MMA. |
|
|
MMA can cut it itself too (MMW does not know actual capacity of the device), also it would not help much in situation when the device is empty and MMA is about to sync thousands of tracks to the empty device, I think MMA should download and analyze track on the same time. I will solve this with Marek as part of 0012138 Re-resolving for testing the original issue 1) |
|
|
Verified original issue in 1724 |
