1 |
P.V.Anthony posted <435D3729.4070507@×××××××××××.sg>, excerpted below, on |
2 |
Tue, 25 Oct 2005 03:34:01 +0800: |
3 |
|
4 |
> I have a problem with my nfs transfer using tcp or udp. There is a pause |
5 |
> every few seconds when I am writing to the server. This causes the video |
6 |
> that I am recording to have jerks and lost frames. The tcp stream is not |
7 |
> constant. |
8 |
> |
9 |
> I will try to explain the best I can. |
10 |
> |
11 |
> I have a server for video use. I connect the videoserver with Macs using |
12 |
> cross cable. Using NFS version 3 and using tcp. The filesystem on the |
13 |
> server is XFS. |
14 |
> |
15 |
> On the macs there is a dv camera connected. Using a video editing |
16 |
> software, I am recording the video to the nfs mount on the mac. After |
17 |
> the recording when I play back the recording there are jerks in the |
18 |
> video every 20 seconds or so. |
19 |
> |
20 |
> Next I copy a big file to the nfs mount. The file goes through but I |
21 |
> notice that |
22 |
> there is pause every 20 seconds or so. DV video only need about 3.5MB |
23 |
> per |
24 |
> second. All the ethernet connection is gigabit. |
25 |
> |
26 |
> What am I doing wrong? Please do advice. |
27 |
|
28 |
Pause every few seconds. I've seen two commonly reported solutions to |
29 |
this -- two different problems resulting in the same symtoms -- and |
30 |
another one less common in general but likely equally common on amd64. |
31 |
|
32 |
1) DMA, or rather, the lack thereof, or interrupt sharing. Ensure that |
33 |
you have the correct chipset drivers installed, both for whatever bus the |
34 |
NIC is on, and for your hard drive interface (the PATA/SATA/SCSI chipset). |
35 |
Often, generic or the wrong drivers will work in degraded mode, but DMA |
36 |
won't be turned on, and performance will suffer. For your hard drives |
37 |
(yes, wrong drivers there /can/ cause bottlenecks up the chain, including |
38 |
onto the same bus as the NIC), use hdparm (or sdparm) to verify that both |
39 |
the drive and the chipset agree that DMA is on. |
40 |
|
41 |
Also check interrupt distribution and sharing. You may be able to change |
42 |
these settings in the BIOS, or try switching some cards around |
43 |
(particularly your NIC if it's not on-board built-in). |
44 |
|
45 |
2) Do you have an "extra" network interface? I've seen this one on |
46 |
MSWormOS and am unsure how Linux would even have the problem, unless you |
47 |
have a spare interface (say a wireless card) that's running but not |
48 |
connected. Anyway, the problem here is an interface set to DHCP mode that |
49 |
can't get an IP because it's not connected, so of course can't see a |
50 |
server. The system would pause every few minutes (every few seconds seems |
51 |
a bit much, but...) and try to obtain an IP, stalling all I/O in the |
52 |
process. (Again, the reason it's stalling all I/O is likely rooted in |
53 |
misconfigured DMA or IRQs, but anyway...) |
54 |
|
55 |
3) Seen frequently on this list as a problem resulting in one or more of |
56 |
a number of symtoms, including this one: APIC or ACPI problems. Take a |
57 |
look at the kernel documentation under $kerneldir/Documentation/x86_64, |
58 |
the apic and acpi boot options, and idle=poll as well. Note that those |
59 |
needing these often find they don't after a mature BIOS is available and |
60 |
flashed. idle=poll is particularly nasty, as it will cause your CPU to |
61 |
work far harder and run far hotter than it ordinarily would, but some |
62 |
early amd64 boards required it until an updated BIOS fixed the issue. I'm |
63 |
not sure if that's still a problem with the newer boards or not. |
64 |
|
65 |
Also note that for specific net chip, various configuration options may be |
66 |
available. In particular, there's the Broadcom tigon3/tg3 kernel driver, |
67 |
and then another (I forget what it's called) for the same net chip. Often |
68 |
one will work well, the other won't work or will have issues. If that's |
69 |
your chip, look it up and consider trying the other driver. |
70 |
|
71 |
-- |
72 |
Duncan - List replies preferred. No HTML msgs. |
73 |
"Every nonfree program has a lord, a master -- |
74 |
and if you use the program, he is your master." Richard Stallman in |
75 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
76 |
|
77 |
|
78 |
-- |
79 |
gentoo-amd64@g.o mailing list |