Re: High CPU Usage transmission-da 2.34 beta 4 (4.1.8)

xushi · ‎2011-09-23

Hi all,

I've seen a few other threads that mentioned this but nothing solid, so I'm starting this thread with the latest Transmission beta and ReadyNAS OS version..

I've been noticing very high CPU usage for transmission-da for the past few months with the past few Transmission versions on my ReadyNAS NV+ RND4000. It reaches a point where it disrupts other services on that NAS. I would like to assist in what further information the devs need, and hopefully for them to take a look at the cause and resolution when possible..

Thanks.

Transmission 2.34 beta 4
ReadyNAS 4.1.8

I also noticed this on the 4.1.8 betas, and other Transmission betas. I cannot recall if this happened in 4.1.7 or not though..

Tasks:  66 total,   5 running,  61 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.0% us,  8.4% sy, 86.1% ni,  1.0% id,  0.0% wa,  0.3% hi,  3.2% si
Mem:   1010560k total,   815584k used,   194976k free,    14576k buffers
Swap:   767904k total,        0k used,   767904k free,   436720k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 2559 nobody    38  13  342m 318m 6816 R 96.4 32.3   1267:41 transmission-da
22570 root      16   0  2992 1680 1328 R  2.6  0.2   0:15.38 top

I just add a few torrents to download (four to five). Even if just 2 are left and downloading at 30% of my bandwidth, it still uses up to 97% cpu usage.

Thanks.

onryo · ‎2011-10-02

I can reproduce the same behavior on an NV+ RND4000 running 4.1.8 and Transmission 2.34b04. Here's my top output:

Tasks:  57 total,   3 running,  54 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.3% us,  8.8% sy, 85.6% ni,  3.9% id,  0.0% wa,  0.0% hi,  0.3% si
Mem:   1010560k total,   380752k used,   629808k free,   203872k buffers
Swap:   767904k total,        0k used,   767904k free,    67312k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                      
 8177 nobody    28   8 27904 8464 5088 R 94.1  0.8   1:53.86 transmission-da                              
 8196 user      16   0  2992 1664 1328 R  2.0  0.2   0:00.52 top

I noticed that it is pegged near 100% when Transmission is running even a single download, but will drop down to 5-10% when paused. When I run a strace on the Transmission process, I see a lot of this in a loop:

writev(56, [{"=W\213\316\334\272\353\246\3028n\254\36&a\240#\330\225"..., 589}], 1) = -1 EAGAIN (Resource 
temporarily unavailable)
gettimeofday({1317576400, 906643}, NULL) = 0
writev(54, [{"$e\273|\211\325\335\230\320\25\2553m\301\253l\0164s\257"..., 556}], 1) = -1 EAGAIN (Resource
͙ temporarily unavailable)
gettimeofday({1317576400, 907946}, NULL) = 0
writev(62, [{"\255\315#H\32q|\267\266\222ul(~J\310\214\233\204\272\17"..., 575}], 1) = -1 EAGAIN (Resource
 temporarily unavailable)gettimeofday({1317576400, 909248}, NULL) = 0
writev(69, [{"\271\241c\233,LD\10\230\312\255+N!\303sN\373\342\274K\334"..., 415}], 1) = -1 EAGAIN (Resour
ce temporarily unavailable)
gettimeofday({1317576400, 910551}, NULL) = 0
writev(91, [{"I\311~b\366\230B\206}h\232\211\321\340`~4\322\213\0054"..., 360}], 1) = -1 EAGAIN (Resource 
temporarily unavailable)

I'm thinking the high CPU is related to the UDP tracker error that others are seeing: http://www.readynas.com/forum/viewtopic.php?f=36&t=56454. Any chance you're having trouble with UDP trackers as well?

MrAsker · ‎2011-10-12

I have basically the same problem but I can't even get anything to DL anymore. 😞

top - 13:29:42 up 11 days, 19:01, 2 users, load average: 1.65, 1.42, 1.25
Tasks: 61 total, 6 running, 55 sleeping, 0 stopped, 0 zombie
Cpu(s): 1.3% us, 1.0% sy, 97.4% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.3% si
Mem: 1010560k total, 523040k used, 487520k free, 159232k buffers
Swap: 511936k total, 704k used, 511232k free, 246608k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20140 nobody 39 18 71680 39m 6160 R 97.7 4.0 14:25.53 transmission-d
19891 root 16 0 2992 1680 1328 R 2.3 0.2 0:30.60 top

Tried various different transmission versions but they all seem to have the same basic problem.

For more info: http://www.readynas.com/forum/viewtopic.php?f=36&t=56454&p=326256#p326256

l3malex · ‎2011-10-14

I confirm, I have also this problem.
Transmission is nearly unusable on my ReadyNas Duo NV

This occurs since my firmware upgrade 4.1.8 or Transmission version 2.34 beta 4

What happen?

odbro · ‎2011-10-15

Same for me..

Totally unusable as it´s now 😞

bino1 · ‎2011-10-17

I raise my hand.
Me too. I have the same pain. Very often connecting to any service of my nas is a bit of slow-pain surgery.
I Hope the developers could find a solution.
I am fully available to help providing information , if i'm able to.

Readynas Duo - Radiator 4.1.8
Transmission 2.34b04

top - 22:16:51 up 16:33,  0 users,  load average: 1.16, 1.18, 1.24
Tasks:  74 total,   2 running,  72 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.0% us,  2.6% sy, 88.0% ni,  7.1% id,  0.2% wa,  0.2% hi,  0.9% si
Mem:    226256k total,   211936k used,    14320k free,    10320k buffers
Swap:   255968k total,   133760k used,   122208k free,    77568k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND            
 1503 nobody    33   8  203m  90m 5968 R 76.0 41.1 898:31.71 transmission-da    
21466 admin     26  10  2976 1536 1216 R  9.5  0.7   0:00.15 top

Has anyone tried to revert back to a previous version of transmission with success?
Has this helped avoid the problem? Thank you

rodolphedj · ‎2011-10-27

Hi,

it is the same thing with transmission 2.42 (last realease) and firmware 4.1.8

top - 13:26:01 up  3:19,  2 users,  load average: 0.99, 0.99, 0.92
Tasks:  64 total,   4 running,  60 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.0% us,  4.2% sy, 94.1% ni,  0.0% id,  0.0% wa,  0.3% hi,  0.3% si
Mem:   1010560k total,   593440k used,   417120k free,    92832k buffers
Swap:   767904k total,        0k used,   767904k free,   449024k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
25043 nobody    33   8 46064  18m 4912 R 98.0  1.8   9:27.94 transmission-da
24913 root      16   0  2992 1664 1328 R  2.0  0.2   0:24.38 top
    1 root      15   0  2000  880  768 S  0.0  0.1   0:05.74 init

Readynas NV+ with 1GB of RAM

transmission configuration :

"cache-size-mb": 300,
"encryption": 2,
"max-peers-global": 200,
"peer-limit-global": 240,
"peer-limit-per-torrent": 40,

false issue :
Ticket #4556 (closed Bug: fixed) https://trac.transmissionbt.com/ticket/4556
available in 2.42 release and cpu usage is always around 100% by transmission processus.

xushi · ‎2011-10-27

Thanks all,

Now I'm sure we all would like a member of staff or developer to look at this thread and assist us.. Alternatively, does anyone know where to go to file a bug report? I could do that and keep the thread updated.

rodolphedj · ‎2011-10-27

for your information, cpu usage is normal now if i disabled :

"dht-enabled": false,
"pex-enabled": false,

top - 14:17:12 up  4:10,  2 users,  load average: 0.49, 0.40, 0.48
Tasks:  62 total,   5 running,  57 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.6% us,  1.6% sy,  2.3% ni, 93.5% id,  0.0% wa,  0.0% hi,  1.0% si
Mem:   1010560k total,   669248k used,   341312k free,    96800k buffers
Swap:   767904k total,        0k used,   767904k free,   518464k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
25551 nobody    23   8 42176  22m 4800 R  3.6  2.3   3:56.20 transmission-da
24913 root      17   0  2992 1664 1328 R  1.6  0.2   1:18.53 top

xushi · ‎2011-10-27

Thanks rodolphedj,

That's useful for a temporary workaround. The problem is some of us do rely on Distributed Hash Tables and Peer Exchanges 😕 I would rather they find the actual cause of the CPU hogging and fix it.

rodolphedj · ‎2011-10-27

it is better when download is slow/null

with lot of torrents and with an usage of high bandwith, the cpu usage is higher, but i think it is normal with encryption job.
↓ 1.03 MB/s ↑ 21 KB/s

you can see the memory usage now is different.

top - 14:29:25 up  4:23,  2 users,  load average: 1.35, 1.10, 0.78
Tasks:  64 total,   4 running,  60 sleeping,   0 stopped,   0 zombie
Cpu(s):  2.3% us, 12.6% sy, 68.6% ni,  0.3% id,  0.0% wa,  3.6% hi, 12.6% si
Mem:   1010560k total,  1002768k used,     7792k free,    11264k buffers
Swap:   767904k total,      320k used,   767584k free,   458528k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
25551 nobody    33   8  505m 485m 4880 R 92.7 49.2  12:57.92 transmission-da
24913 root      16   0  2992 1664 1328 R  2.9  0.2   1:33.37 top
25681 media     15   0 10544 4816 3440 S  2.3  0.5   0:10.20 smbd

ObiWan84 · ‎2011-10-27

Where did you get 2.42 from? Did you compile it yourself? If so you may want to try compiling it with --enable-lightweight option as see if this improves things.

rodolphedj · ‎2011-10-27

ObiWan84 wrote:
Where did you get 2.42 from? Did you compile it yourself? If so you may want to try compiling it with --enable-lightweight option as see if this improves things.

Hi,

Did you compile it yourself = yes (viewtopic.php?f=36&t=58331)
--enable-lightweigh = No improvement, because it is only a default parameter apply about encrypted set at without.
And in my personal configuration, I've forced the data encryption at mandatory.