Indexing Stuck + High Cpu Usage

[David]

Hello !

 

I've been testing btsync for a few days and I was able to sync folders containing a few files/folders succesfully. Btsync seems to be the solution I was waiting for.

 

Now I'm running into a problem : I'm trying to sync a folder containing about 35000 files (about 20Gb) between two synology NAS boxes so I added a folder share on the first one and generated a secret.

 

The btsync web interface tells me this :

 

21.7 GB in 36915 files (Indexing...)

 

Also, the CPU usage is abnormally high (above 70%) and I'm talking about constant usage, not peak usage.

 

My folder contained originally 500k+ files and I thought that it was too many, that's why I reduced the number of files to about 35k.

 

Indexing seems to be stuck, could someone help me troubleshoot this ? I've looked into the logs but I could not find anything relevant.

 

Thanks in advance,

 

David

Edited November 29, 2013 by David

[Harold Feit]

What version of Sync?

[David]

Version is 1.2.73

Edited November 29, 2013 by David

[Harold Feit]

Posting the logs (zip them if necessary) should provide information that we'll be able to identify the cause with

[David]

Hi,

 

I had the debug option turned on and my log file grew up to 3Gb... I've cleared it and started over. Do you want the logs with the debug option on or off ?

 

In the meantime, here's the content of my newly created log file right after I start btsync :

[20131202 15:12:45.440] total physical memory -1 max disk cache 2097152[20131202 15:12:45.442] Using IP address 192.168.0.17[20131202 15:12:45.444] test sha1: AE5BD8EFEA5322C4D9986D06680A781392F9A642[20131202 15:12:45.444] test sha2: 630DCD2966C4336691125448BBB25B4FF412A49C732DB2C8ABC1B8581BD710DD[20131202 15:12:45.444] test aes: 0A940BB5416EF045F1C39458C653EA5A07FEEF74E1D5036E900EEE118E949293[20131202 15:12:45.445] Loading config file version 1.2.73[20131202 15:12:57.466] NAT-PMP: Unable to map port with NAT-PMP.

At the moment the btsync process is running at about 80% CPU. The web interface isn't showing any folder nor the version. The "Add Folder" and "Preferences" buttons aren't responsive.

 

Edit :

 

Now the web interface is showing the folder and is responsive.

 

Here's an extract from the log file :

[20131202 15:12:57.466] NAT-PMP: Unable to map port with NAT-PMP.                                                                                                                                                           6[20131202 15:22:50.784] SyncFolderScanner: Adding notifer "/volume1/Backup_Data"[20131202 15:22:50.798] Added path=/volume1/Backup_Data with fd=1[20131202 15:22:50.908] Added path=/volume1/Backup_Data/ApplicationsInternes with fd=2[20131202 15:22:51.018] Added path=/volume1/Backup_Data/Developpement with fd=3                                                                                                                          2[20131202 15:22:51.185] Added path=/volume1/Backup_Data/Administratif with fd=4

and it goes on and on with paths.

[nils]

1. It seems, that BTsync is still indexing - it might take a while, depending on the NAS's CPU type (ARM vs x86).

2. the debug option might be able to help you, but I would suggest letting it run for a long while and then see if it is still stuck e.g. 2-6h and have a look at the log file at that point

3. use the latest version 1.2.82

[David]

Ok it's been running for about 20 hours, the web interface still says it's indexing.

 

Here's an extract from the logs (I'm obfuscating file names) :

[20131203 11:09:59.959] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/aaaa[20131203 11:09:59.959] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/bbbb[20131203 11:09:59.960] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/cccc[20131203 11:09:59.961] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/dddd[20131203 11:09:59.961] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/eeee[20131203 11:09:59.962] SyncFilesController [file updated]: processing file /volume1/Backup_Data/Administratif/ffff

[David]

Hi,

 

Btsync has been running for days now, and is still indexing. My logs are full of "file updated" statements, but I know these files aren't modified. I can see such statements occuring for the same file over and over again :

RackStation> cat sync.log | grep /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext[20131204 00:01:46.675] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 00:11:45.409] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 00:21:47.084] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 00:31:48.998] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 00:41:45.034] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 00:51:47.401] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:01:47.413] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:11:45.684] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:21:46.918] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:31:46.583] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:41:44.974] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 01:51:47.096] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:01:46.943] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:11:46.102] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:21:45.826] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:31:46.735] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:41:44.769] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516[20131204 02:51:48.337] SyncFilesController [file updated]: processing file /volume1/Backup_Data/ApplicationsInternes/fileabcd.ext 1266236920 516

And this happens for all the files actually.

 

First of all, does "file updated" mean that btsync thinks the file has been modified ? If so, how does it determine that ?

 

Any help would be much appreciated.

 

Thanks,

 

David

[knaapie]

Hello,

 

I have the same issue, so I was wondering if anyone has some comment on the OP's issue?

 

I have just set up the sync folder, so it's setting up a first index. Now it's stuck on 36.8GB and 156700 files. And showing 'SyncFilesController [file updated]:' messages, but not adding new torrents.

 

This is the second time I added this sync folder, I recreated it with a new secret to see what would happen.

First time it was stuck on 36.8GB and 156700 files as well.

 

I'm running BTsync 1.2.82 on OS X 10.6.8 (Snow Leopard)

The files I'm indexing are actually on a shared drive on my network. That may be an issue?

[David]

Hi,

 

The issue isn't fixed yet, I just can't use btsync...

 

I've e-mailed sync@bittorrent.com but haven't got any reply.

 

Regards,

 

David

[goli]

Hey there.

 

Did you mount your file system without file creation time or file modification time?

 

Regards,

Stephan.

[David]

Hi Stephan,

 

In my case, both are default ext4 filesystems created on 2 synology NAS, here're the mount options (data to be synced is located on /dev/mapper/vol1-origin) :

RackStation> mount/dev/root on / type ext4 (rw,relatime,barrier=0,journal_checksum,data=ordered)/tmp on /tmp type tmpfs (0)none on /dev/pts type devpts (gid=4,mode=620)/sys on /sys type sysfs (0)/proc/bus/usb on /proc/bus/usb type usbfs (0)/dev/mapper/vol1-origin on /volume1 type ext4 (usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,synoacl)

Regards,

 

David

[wiwiec]

I'm having the same issue with version 1.3.94.

 

I have btsync running on three machines, 2 x linux_x64 and 1 x raspberry pi arm.

 

The CPU load on the linux machines seem to be reasonable after peak time. The raspberry pi however has a constantly high load of 70-80%. 

 

The sync.log shows these lines, no output for a long time after that (while high CPU load)

[20140518 01:14:34.795] Incoming connection from *****:34477[20140518 01:14:35.911] Got id message from peer ***** - ***** (*****) 1.3.94[20140518 01:14:35.912] Got state sync request from peer *****[20140518 01:14:36.169] Merge: processing get_root message, my hash: *****[20140518 01:14:37.858] Merge: processing get_have_pieces message, my hash: *****, prev hash: *****[20140518 01:14:39.235] State sync finished for folder /media/GIGABEAT_2/btsync

My synced folder resides on a USB connected harddrive and is mounted as:

/dev/sdb2 on /media/GIGABEAT_2 type ext4 (rw,nosuid,nodev,uhelper=udisks)

Ext4 itself should not be the problem, as my linux machines also sync to ext4 filesystems.

The mount flags might be a problem?

[RomanZ]

@wiwiec,

 

Does it happens when actual sync is going or even when no sync occurs? It would be normal when sync is transferring data: it has to hash all the data it gets (while hashing is rather heavy operation for low-power CPU).

[feanorknd]

Hi David...

 

I am not experienced and have own problems...... but check this:

 

root@ip-10-0-1-10:/var/lib/btsync# cat /proc/sys/fs/inotify/max_user_watches
1048576

 

The inotify max number should exceed by far the number of files synced.. I guess... I have had a problem due to this I think, and it was translated at very high CPU.

 

 

Hi,

 

The issue isn't fixed yet, I just can't use btsync...

 

I've e-mailed sync@bittorrent.com but haven't got any reply.

 

Regards,

 

David

[David]

Hi David...

 

Hi feanorknd,

 

Thanks for your input, but unfortunately I've abandoned BitTorrent Sync for many reasons, one being lack of support.

 

I've had 2 big issues which were never resolved so I couldn't go further with it.

 

Regards,

 

David

[bradbamford]

I'm having the same problem with a folder that is about 1.8TB

The client has been hung for 1 month just saying indexing 

[Emosan]

Same here with version 2.0.93, 200Gb and more than 10 million files, with two ARM NAS 

[kyamon]

I am also having this problem, or more generally that Sync just stops before finishing the first sync.

My setup: iMac at home, MBPro in office, RaspberryPi w/ ext HD at home.

MBP in the office is always running, in order to correctly sync with the iMac; but I would like to change that, hence the RB.

Sync on MBP is 1.4, the others are 2.0; in total about 35-40 GB to sync.

 

What happened is that the RB started syncing and moved to about 28 GB but then stopped. It now very often shows that it is indexing, but it does not sync the remaining files, and it uses 100% CPU. If I create an empty file on the RB (using "echo>foo") then that gets synced through to the others, but if I create a file on the iMac then syncing does not happen, or only after a very long time.

 

I read in some other post that filenames that contain special characters can be a problem, and I have many of those in Keynote presentations, as well in other Software-generated files, so I can not change that. Is this issue confirmed, and could it be a reason for my problems? In my case the files that don't get synced also have normal filenames, but maybe there are so many others that sync gets confused...
I did a test regarding that and made a second sync folder that only contains my Keynote presentations. It is in total about 3.2 GB on my iMac, but only 2.2 get synced to the RBPi (in this case it does not sync to the MBP, so only Sync 2.0 devices are involved). When I checked progress during the first sync I noticed that sync basically works down the file/folder list alphabetically. I have several subfolders for the past 10 years, and they get synced starting with the lowest numbers. The syncing stops at 2014 (of course...   ).

 

Sync worked flawlessly between the two macs, but not with the RBPi - I am worried that this will scramble my files somehow, and that the Pi will make Sync delete files or something. Does anybody have an idea what is going on here?

[maximbroos]

I still have the same issue... Is there any progress on this matter?

My raspberry pi just keeps trying to sync a read-only folder (located on a attached flash drive), but after syncing 13 GB, it keeps on indexing forever (for three days now).

While indexing it also munches up all the available CPU time, rendering the raspberry almost useless,  so i'm about to look for another solution. 

Any support would be greatly appreciated!

[Innomen]

Same with me. Windows 7 two way to android 4.2.2. Two folders 25gb total, photos. It just indexes forever on the phone side. Two days now. 530 ram and around 50% cpu usage.

 

If this kills my sd card I'm going to be sadface and madface

 

Creating support ticket.