Problem copying files to Bubba

[MartinHageras]

Hi!

I have trouble copying large number of files to my Bubba Two. I have tried copying with explorer and the Web File manager in bubba.

Using the explorer it works for a while, but it seems rather slow. Also I get logged out after a while, which interrupts the copy process.

Using the Upload in the File Manager doesn't seem to work at all. It never leaves the "Initializing" stage.

What is the best way to copy files to bubba? Should I copy from my local disks or is it better to copy from my network share to bubba?

Regards, Martin.

EDIT:
I'm trying to copy about 5000 files totalling 30 GB or so.

[Tompa]

I've had similar problems when I copied lots of data at the same time using samba (explorer). After a while I recieved a "Disk full?" error but there was lots of space left.

The problems went away when I copied through FTP instead. I haven't tried on the latest firmware though.

Edit: Spelling

/Tomas

[PaulW]

I have the same problem too. The initialization stage remains and nothing happens anymore. It was just a images from approx 4 GB of size......
The software version I use is Version 1.0.8-1

When I use FTP then everything works fine. But using the browser (In this case firefox) the problem as described above occurs.



/PaulW

[ahab]

I have basicly the same problem... I have mailed the excito support but still no answer.

http://forum.excito.net/viewtopic.php?t ... 238dedde96

Haven't even got ftp to work.

[johannes]

Hi all,

This is new to us, we are trying our best to reproduce it in order to find the issue.

Do I understand correctly that this is two separate issues?

1) Bubba goes down when copying large amounts of data?

2) Filemanager upload doesn't handle large files. This is a known issue we are working with, if I remember correctly the limit is 2 GB.


Any details or other ways to reproduce 1) are much appreciated. I have copying files all day but haven't been able to reproduce yet.

[PaulW]

Hi Johannes,

From my point of view the message from ahab describes a different problem. My bubba is not going down. It simply does not react anymore after transferring a bigger file.
So my issue is the same as explained in the first message from MartinHageras.
So you are right that we have two different issues here.

Kind regards,

Paul

[ahab]

Hi Johannes,

I guess it's not easy being a support guy when you have to deal with technology idiots like me... I'm not sure if we're talking about different things or just using different words to describe the same thing.

However, this is my interpretation:

The following problems seems to exist:

1. Problem when using the upload manager - it cannot handle single files that are too big nor can it handle lots of small files that totals to a lot. This is one problem I haven't experienced (haven't tried the upload manager).
2. Problem with disk size when using the explorer (samba), problem solved by using ftp
3. Problem when copying/moving lots of files via explorer (samba?). It's slow and after a while the bubba stops responding. This is my problem. I'm not sure if there's difference between "stop responding" and "goes down". For me the bubba disappears from the network/router.

I have two questions for Martin to see if our problem with the explorer is the same;

1. Have you tried using ftp? For me that doesn't work either. It's just as slow and interrupts after a while and then I cannot connect to the bubba
2. Is your bubba accessible after the transfer has interrupted? I have to reboot it.


A question to all you guys; how's your network setup? I'm using a netgear router/firewall as my internet connection and the bubba is only accessible from inside my network.

As for Johannes question. If you could specify what kind of information you need for troubleshooting I'd be glad to help you as much as I can. You have my phonenumber from my mail to support@exito.com. Please, feel free to give me a call.

[Tompa]

1) Bubba goes down when copying large amounts of data?

I'm not sure how to reproduce the error, but here's an extract from my samba-log, and from what I can see some problems occur there:

Code: Select all

[2008/10/21 17:51:40, 1] smbd/service.c:make_connection_snum(950)
dimensionxp (192.168.1.8) connect to service home initially as user xxxxx (uid=1001, gid=100) (pid 32699)
[2008/10/21 17:51:40, 1] smbd/service.c:close_cnum(1150)
dimensionxp (192.168.1.8) closed connection to service home
[2008/10/21 17:51:40, 1] smbd/service.c:make_connection_snum(950)
dimensionxp (192.168.1.8) connect to service home initially as user xxxxx (uid=1001, gid=100) (pid 32699)
[2008/10/21 18:26:53, 0] lib/util_sock.c:write_data(562)
write_data: write failure in writing to client 192.168.1.8. Error Connection reset by peer
[2008/10/21 18:26:53, 0] lib/util_sock.c:send_smb(761)
Error writing 4 bytes to client. -1. (Connection reset by peer)
[2008/10/21 19:05:03, 1] smbd/service.c:make_connection_snum(950)
192.168.1.5 (192.168.1.5) connect to service home initially as user xxxxx (uid=1001, gid=100) (pid 441)
[2008/10/21 19:51:58, 0] lib/util_sock.c:read_data(534)
read_data: read failure for 4 bytes to client 192.168.1.8. Error = No route to host
[2008/10/21 19:51:58, 1] smbd/service.c:close_cnum(1150)
dimensionxp (192.168.1.8) closed connection to service home
[2008/10/21 21:38:02, 0] lib/util_sock.c:write_data(562)
write_data: write failure in writing to client 192.168.1.8. Error Connection reset by peer
[2008/10/21 21:38:02, 0] lib/util_sock.c:send_smb(761)
Error writing 4 bytes to client. -1. (Connection reset by peer)
[2008/10/21 21:38:02, 0] lib/util_sock.c:get_peer_addr(1221)
getpeername failed. Error was Transport endpoint is not connected
[2008/10/21 21:38:02, 0] lib/util_sock.c:get_peer_addr(1221)
getpeername failed. Error was Transport endpoint is not connected
[2008/10/21 21:38:02, 0] lib/util_sock.c:write_data(562)
write_data: write failure in writing to client 0.0.0.0. Error Connection reset by peer
[2008/10/21 21:38:02, 0] lib/util_sock.c:send_smb(761)
Error writing 4 bytes to client. -1. (Connection reset by peer)
[2008/10/21 22:03:03, 1] smbd/service.c:make_connection_snum(950)
dimensionxp (192.168.1.8) connect to service home initially as user xxxxx (uid=1001, gid=100) (pid 1006)
[2008/10/22 00:09:44, 1] smbd/service.c:close_cnum(1150)
dimensionxp (192.168.1.8) closed connection to service home

Besides the problem that is "goes down" after a big amount of data has been copied it also seem to act slow and heavy when you map bubba2 as a network drive and try to reach it for the first time. It freezes the whole "My computer" dialog and 20-30 seconds later it starts to respond again. I don't know if this is related to the copy issues described above.

The "mapping" problem seem to have increased in the latest firmware (?) and happens almost every time for me as I have a default mapping script that runs when I login to my computer.

My setup is that bubba2 is acting router/firewall on the wan port and the rest of the computers are on the lan port.

/Tompa

[CJNE]

Happens for me as well, trying to copy 194 Gb via samba.
The file transfer can go on for quite a while (at least an hour or so) before it fails and bubba locks up.
It still responds to ping and the router functionality is still working.
I was logged in (as root) when this happened and i can't do anything in the shell. No kernel messages appeared.
Power down also doesn't work so i guess i'll have to pull the plug...
Seems to be some kind of resource exhaustion by the way it behaves.

I managed to get the same amount of data onto my bubba2 earlier using FTP, haven't tried with this firmware version though.

I'm also a bit unimpressed with the transfer rate, i get somewhere around 6-9 MB/sec on a gigabit connection, is this normal?
Using rsync or scp only gives my 2 MB/sec, ftp around 5 MB/sec.

[CJNE]

I left it running and it actually restarted after 6 hours. Can't find anything interesting in the logs...

[Jockewhbg]

I think I have the same problem. I wrote about this before. I connected my Bubba as a network-drive in Explorer(Samba). Then I used it to save downloads from my Torrent-client. When my Torrents had a high speed, say around 2 Mbyte/s, my Bubba didn´t last long. After approx one hour it slowed down a lot. Downloads decreased to around 20Kb/s and the Torrent client complained about disk overload. I shut down the Torrent-client and tried to run other things from the Bubba but everything was very slow, browsing, music was not possible, watching video was not possible. Only a restart of the Bubba solved the problem. It seems that some kind of overload occured. I solved the problem by decreasing my download speed to about 1,2Mbyte/s and then it worked fine.

I can imagine that others would have problem when copying a lot of files over samba where regulating the transfer rate isn´t possible.

Regards

Jocke

[amry]

What I read is that most users experienced bubba's passing away using samba. In my opinion it can´t be related to samba only. After copying about 5GB (2 x 2GB + 1GB files) via ssh my bubba died as well. Wasn't able to conduct proper shutdown (via webinterface, cli or the button on bubba 's backside). I had to pull the plug. I checked dmesg after reboot and the only strange thing I discovered (I am not a linux expert, so bare with me) was :

Code: Select all

EXT3-fs: INFO: recovery required on readonly filesystem.
EXT3-fs: write access will be enabled during recovery.
kjournald starting.  Commit interval 5 seconds
EXT3-fs: sda1: orphan cleanup on readonly fs
ext3_orphan_cleanup: deleting unreferenced inode 32653
ext3_orphan_cleanup: deleting unreferenced inode 32652
ext3_orphan_cleanup: deleting unreferenced inode 687271
ext3_orphan_cleanup: deleting unreferenced inode 687254
ext3_orphan_cleanup: deleting unreferenced inode 687253
ext3_orphan_cleanup: deleting unreferenced inode 32650
ext3_orphan_cleanup: deleting unreferenced inode 32649
ext3_orphan_cleanup: deleting unreferenced inode 32648
ext3_orphan_cleanup: deleting unreferenced inode 32647
ext3_orphan_cleanup: deleting unreferenced inode 32646
ext3_orphan_cleanup: deleting unreferenced inode 32645
ext3_orphan_cleanup: deleting unreferenced inode 32644
EXT3-fs: sda1: 12 orphan inodes deleted
EXT3-fs: recovery complete.
EXT3-fs: mounted filesystem with ordered data mode.

but my assumption is that's a consequence of pulling the plug instead of proper shutdown.

[jws]

Hi all,

This is new to us, we are trying our best to reproduce it in order to find the issue.

Do I understand correctly that this is two separate issues?

1) Bubba goes down when copying large amounts of data?

2) Filemanager upload doesn't handle large files. This is a known issue we are working with, if I remember correctly the limit is 2 GB.


Any details or other ways to reproduce 1) are much appreciated. I have copying files all day but haven't been able to reproduce yet.

In my case I mounted the "music" directory of my Debian Sid machine onto bubba through nfs, then did

Code: Select all

time cp /mnt/music /* /home/storage/music

("time" because I wanted to know how long it took, in order to calculate the transfer speed).

After about 10 minutes the bubba froze; shell logins became unresponsive; ping to the bubba did not return anything ("no route to host"). The reset button on bubba was also frozen, didn't do anything. Had to pull the plug

After restarting the bubba, it turned out that about 5.5G (out of 9G) had been copied, until some filenames beginning with M. To copy from M to Z, I used the -u option on cp:

Code: Select all

time cp -u /mnt/music /* /home/storage/music

which skips the files that already have been copied (i.e. for which the source time is not newer than the destination time). Took 9 minutes. So approx. 100 Mbits (not Mbytes) per second. I blame my network switch, because the bubba (I believe) and my PC can do 1000 Mbits/sec. Well, if so, it wasn't so bad really.

During that second copying prodedure I kept a login shell open with "top" in it. The CPU use of cp was a bit alarming; mostly close to 80%, sometimes higher. Could the bubba have crashed because of some limit (perhaps a temperature limit?) which was exceeded during the first attempt?

Anyway. I now think that the "copying large amounts" problem may not be caused by a hard limit on the size of the cp operation. It is perhaps more a question of making it behave more "nicely" (see "man nice". Of course on the bubba2, "man" is almost unusable until you install "less" -- a strange omission).

[johannes]

Ok, thanks all for your feedback. We are working hard on this one, since this kind of crashing is quite worrying for a server. Hope to have some update for you all soon.

[jws]

Ok, thanks all for your feedback. We are working hard on this one, since this kind of crashing is quite worrying for a server. Hope to have some update for you all soon.

I hope so too, because I am still within my 30-day return period, and if it is not fixed the Bubba2 has to be called "unusable", and should go back to the shop.

Latest results:

I mounted an external usb disk, and attempted to transfer some soap opera episodes to /home/storage/video.

1) "cp -a /mnt/MySoap /home/storage/video" failed almost at once. Freeze; non-response; only plugout - plugin helps. Even worse, the home internet connection (for which the bubba 2 is the gateway) was cut (including my wife's Skype conversation with our daughter).

2) "nice cp -a /mnt/MySoap /home/storage/video" failed after copying one episode (about 700 MB) and half of another one. "top" showed almost constant 74% CPU use for the cp operation.

3)After plugout/plugin/restart I tried to copy the episodes one by one, instead of in a bunch. By this time the bubba 2 was so "overheated" that this also failed, almost at once.

So really, this bubba2 is almost on its way back (I am glad I kept the packaging). Was this 1000 Mb/s transfer rate that Excito puts in its advertising ever tested for any length of time, I wonder.