ubuntuone-syncdaemon STILL consumes ridiculous amounts of ram and grinds my machines to a halt

[Jeff Lane ]

I've been reporting this since Lucid. In 4 releases now, I have NEVER been able to reliably use Ubuntu One, despite numerous attempts at giving it a shot. In Natty, I was at the point where I was ready to purge Ubuntu One completely from my systems because ubuntuone-syncdaemon consistently makes my machines unusable. Instead of just purging it, I thought I'd give it another shot with Oneiric.

Since upgradeing my systems to Oneiric, the problem is still present. Unfortunately, this impacts my ability to work, as the syncdaemon consistently eats up at least 50% of available ram and up to 100% of CPU resources, effectively DoSing my machines. :(

I have three systems that are currently running Oneiric.

1: Netbook w/ an older Atom and 1.5GB RAM
2: Lenovo x201 w/ core i5 and 4GB RAM
3: Alienware M15x w/ core i7 and 4GB RAM.

On all three, allowing syncdaemon to run causes the machines to become unusable as the syncdaemon thrashes my hard disks, consumes pretty much all CPU (or on my multi-core systems, it pegs at least one core at 100%) and consumes 50 - 90% system RAM.

On my Alienware, I've already deleted all ubuntu-one startup files to prevent Ubuntu One from running.
Same on my netbook.

The only system I have left now, after 4 attempts at using U1 is the x201, and I fear I now need to purge U1 from it as well, as I need this machine to work when I'm travelling, and I can't work while U1 is thrashing the system.

Here's a snip from top that I managed to get:
top - 10:55:25 up 1:18, 0 users, load average: 1.76, 3.18, 3.47
Tasks: 218 total, 3 running, 210 sleeping, 0 stopped, 5 zombie
Cpu(s): 28.0%us, 2.1%sy, 0.2%ni, 56.2%id, 13.4%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 3719004k total, 3603696k used, 115308k free, 7532k buffers
Swap: 1998844k total, 683212k used, 1315632k free, 334028k cached

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 1892 bladernr 20 0 2440m 1.9g 6584 R 97 53.2 69:25.06 ubuntuone-syncd
   31 root 20 0 0 0 0 D 2 0.0 0:13.20 kswapd0
 1945 bladernr 20 0 1320m 63m 10m S 2 1.8 2:38.47 chrome
 5208 bladernr 20 0 927m 45m 8088 S 2 1.3 0:21.01 chrome

notice that syncdaemon is consuming 97% CPU time (on one core, thankfully, not across all cores) and 1.9GB of RAM out of 4 in the system. It's in a "Local Scan" so in addition to consuming a significant portion of RAM and CPU, it's also doing a LOT of disk I/O which further slows the system down. I could let this sit for HOURS and the result would be the same.

Here's the current status:
bladernr@GarbageScow:~$ u1sdtool -sCreated new window in existing browser session.

State: LOCAL_RESCAN
    connection: Not User With Network
    description: doing local rescan
    is_connected: False
    is_error: False
    is_online: False
    queues: WORKING

ProblemType: Bug
DistroRelease: Ubuntu 11.10
Package: ubuntuone-client 2.0.0-0ubuntu2.2
ProcVersionSignature: Ubuntu 3.0.0-13.22-generic 3.0.6
Uname: Linux 3.0.0-13-generic x86_64
ApportVersion: 1.23-0ubuntu4
Architecture: amd64
Date: Wed Dec 14 11:00:32 2011
InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Release amd64 (20111012)
PackageArchitecture: all
SourcePackage: ubuntuone-client
UpgradeStatus: No upgrade log present (probably fresh install)

[Jeff Lane  (bladernr)]

I would love to be able to see what is in queue, but unfortunately, this happens:

bladernr@GarbageScow:~$ u1sdtool --waiting

Oops, an error ocurred:
Traceback (most recent call last):
Failure: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

So maybe I'll at least look and see what folders U1 is scanning:

bladernr@GarbageScow:~$ u1sdtool --list-folders

Oops, an error ocurred:
Traceback (most recent call last):
Failure: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

I may as well just shut it down now so I can actually get some work done:

bladernr@GarbageScow:~$ u1sdtool -q
ubuntuone-syncdaemon stopped.

Surprisingly, this actually worked this time. about 50% of the time, syncdaemon is consuming so much resources that the quit command actually fails. Instead of the above message, I get a message like this:

ubuntuone-syncdaemon still running.

Look, I'm still, after 2 years, willing to help make this work, and I'm even willing to conceded that I could possibly be part of the cause (pebkac) however, I find it hard to believe that I am the direct cause of exactly the same symptoms on 3 different machines.

[Jeff Lane  (bladernr)]

So after killing the syncdaemon, I re-ran list-folders and actually got a list of the folders on this machine that U1 is supposed to sync:

bladernr@GarbageScow:~$ u1sdtool --list-folders
Folder list:
  id=d6993e0a-b45d-4285-8b47-373862b3530b subscribed=True path=/home/bladernr/Documents
  id=ad119263-c621-40e0-9d8c-b4e425e30afa subscribed=True path=/home/bladernr/development
  id=c75d1395-ee3a-49f8-8515-26219f6d15b4 subscribed=True path=/home/bladernr/Datacenters
  id=767e6419-9061-40b2-9a2e-3fed5440b276 subscribed=True path=/home/bladernr/cia
  id=d65bec91-06b7-40b4-a813-7772014599cd subscribed=False path=/home/bladernr/.ubuntuone/Purchased from Ubuntu One
  id=6ec1aef5-87c5-4620-a846-8bae2dbd23cb subscribed=True path=/home/bladernr/Music

bladernr@GarbageScow:~$ du -hsc Documents development/ Datacenters/ cia/ Music/
13G Documents
5.7G development/
324M Datacenters/
14M cia/
2.6G Music/
22G total

This machine does not have a .ubuntone folder... but it's also not subscribed so that shouldn't matter.

[Jeff Lane  (bladernr)]

bladernr@GarbageScow:~/.config$ u1sdtool -q
ubuntuone-syncdaemon still running.

There ya go... it's currently using up 2.4 of 4GB RAM and won't die :(

[Leo Arias (elopio)]

Hi Jeff,

We are tracking the debugging and progress of memory consumption on a master bug. I've marked this as a duplicate. That way you will be able to follow the work that's being done.

Thanks for your report.

[Jeff Lane  (bladernr)]

Thanks Leo... I'm now following the master for this :)