I’ve been googling the world wide web and more and cannot find an answer to my problem. It seems that there is a hard 1 MB/s download limit for the whole 12.1 system I have installed. All other internett connected equipment I have can fully utulise my 60Mb/s internett line. So, do OpenSUSE have a hard wired bandwith limit? And how on earth to I get rid of it?
Well, my ISP is giving me 60MB/s guaranteed, usually more, on all my other devices.It is just my OpenSUSE install that is so painstakingly slow on downloads.
On Mon, 04 Jun 2012 21:16:02 +0000, jooiong wrote:
> So, do OpenSUSE have a hard wired bandwith limit?
No. Such a limit would be ridiculous to impose and there would be no
reason for doing it.
So help us out here - how are you determining that you’re only seeing 1
MB/s download speed - what site are you connecting to, for example?
It may well be that the limitation is at the other end of the connection
either a physical limitation or some bandwidth-throttling being done
from the site you’re attempting to download (some sites do that so one
person with a huge pipe can’t monopolize the bandwidth).
On Mon, 04 Jun 2012 21:26:02 +0000, jooiong wrote:
> Well, my ISP is giving me 60MB/s guaranteed, usually more, on all my
> other devices.It is just my OpenSUSE install that is so painstakingly
> slow on downloads.
So again, tell us where you’re trying to download from.
You’ve lept to a conclusion as to what the problem is rather than
describing the problem in a way that people can assist you.
My location is Norway and my ISP is Altibox. Fiber to the home with a 60/60 line.
Now, often I download simultaneously from several sites. My Fedora 16 installs, my kids Macs, iPads and Galaxy S2 phones all download at “full” speed. This also go for my PS3, Samsung SmartTV and other gadgets. It is only on my OpenSUSE box that I never ever go beyond the 1MB/s “limit”. No matter from where, when or how I download. Downloading from several different sites at the same time just slows down all downloads so the aggregated bandwidth use is 1MB/s. I can simultaneously download on my Fedora 16 laptop at the reamaining bandwidth (approx 6MB/s) from the same sites. Thus making me point my finger to OpenSUSE it self.
So, where to start trouble shooting? Do there exist anything I can have turned on by mistake that will throttle bandwith? QoS or someting?
On Mon, 04 Jun 2012 21:56:02 +0000, jooiong wrote:
> So, where to start trouble shooting? Do there exist anything I can have
> turned on by mistake that will throttle bandwith? QoS or someting?
Not installed by default. What is the physical network hardware in the
system?
It might be useful to do the following:
Reboot the system
Attempt a download
After the download has been running for a little bit, grab the output
of: uptime && netstat -s (this gives the uptime of the system and some
network statistical counters - the time up is important so we can look at
the stats in the context of how long the system has been up and with a
single file transfer attempting to be run).
It might also be useful to see if you can achieve a higher transfer rate
on your local network, say from your fedora system.
That will tell us if it’s an interaction between the system and your
router (maybe an MTU setting that needs to be tweaked, for example).
> I do not think my HW is to blame:
We’ll see. It could be a hardware-related issue. Have you tried a LiveCD
(openSUSE and/or another distro) to see if you can get a faster download
with that? If you can try that, that’ll tell us if it’s the install on
the system or something else.
Does your ISP support IPv6? If it doesn’t, you might try disabling it
(though I don’t think that would slow performance down as you describe -
it would slow down name resolution if you had to wait for a timeout for
IPv6 resolution, but once a transfer starts, the system doesn’t need to
resolve the address again).
What is the output of a similar ifconfig command on your Fedora box? (If
the MTU is having to be renegotiated constantly, that might turn up as a
difference between the two systems)
On Mon, 04 Jun 2012 22:09:45 +0000, Jim Henderson wrote:
> 1. Reboot the system
> 2. Attempt a download
> 3. After the download has been running for a little bit, grab the
> output of: uptime && netstat -s (this gives the uptime of the system
> and some network statistical counters - the time up is important so we
> can look at the stats in the context of how long the system has been up
> and with a single file transfer attempting to be run).
Additional note, on step 3, when you post that output, be sure to include
it in code tags so it isn’t wrapped inappropriately - otherwise it won’t
be readable.
You’ll need to be in the advanced message editor to do that.
On Mon, 04 Jun 2012 22:58:07 +0000, Carlos E. R. wrote:
> How is your ethernet configured? Use “mii-diag”.
> Are you sure your ethernet cable is rated for 100Mbps?
The number of errors suggests that it’s not a cable quality problem - if
it were and it negotiated 100 Mbps without an issue, then the error
counter would be going nuts.
On 2012-06-05 01:44, Jim Henderson wrote:
> On Mon, 04 Jun 2012 22:58:07 +0000, Carlos E. R. wrote:
>
>> How is your ethernet configured? Use “mii-diag”.
>> Are you sure your ethernet cable is rated for 100Mbps?
>
> The number of errors suggests that it’s not a cable quality problem - if
> it were and it negotiated 100 Mbps without an issue, then the error
> counter would be going nuts.
It could have negotiated 10 Mbps.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
On Tue, 05 Jun 2012 00:23:06 +0000, Carlos E. R. wrote:
> On 2012-06-05 01:44, Jim Henderson wrote:
>> On Mon, 04 Jun 2012 22:58:07 +0000, Carlos E. R. wrote:
>>
>>> How is your ethernet configured? Use “mii-diag”.
>>> Are you sure your ethernet cable is rated for 100Mbps?
>>
>> The number of errors suggests that it’s not a cable quality problem -
>> if it were and it negotiated 100 Mbps without an issue, then the error
>> counter would be going nuts.
>
> It could have negotiated 10 Mbps.
Now there’s a thought I hadn’t considered - that would be 1.25 MB/s max
throughput, take away overhead, that would be about 1 MB/s max speed.
On 2012-06-05 02:35, Jim Henderson wrote:
> On Tue, 05 Jun 2012 00:23:06 +0000, Carlos E. R. wrote:
>> It could have negotiated 10 Mbps.
>
> Now there’s a thought I hadn’t considered - that would be 1.25 MB/s max
> throughput, take away overhead, that would be about 1 MB/s max speed.
precisely…
And it happened to me on a site where they were using correct ethernet
cable, that tested ok statically, pins 1 to 1, 2 to 2,… 8 to 8. BUT! the
pairs were incorrect, 3 and 5 were not a twisted pair, they had paired 1-2,
3-4, 5-6, 7-8. Just fine for telephony, plain wrong for ethernet at 100 mbps.
You get two different results on that case: a) negotiation fails and go
down a step. b) it does not fail, for there are a lot of transmission
errors. I think it depends on the length of cable what case you get.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Carlos E. R. wrote:
> On 2012-06-05 02:35, Jim Henderson wrote:
>> On Tue, 05 Jun 2012 00:23:06 +0000, Carlos E. R. wrote:
>
>>> It could have negotiated 10 Mbps.
>> Now there’s a thought I hadn’t considered - that would be 1.25 MB/s max
>> throughput, take away overhead, that would be about 1 MB/s max speed.
>
> precisely…
Right. So why has nobody suggested that the OP post the link speed?
jooiong please post the output (in CODE tags) of: