All of my network cables have been from monoprice. However the last batch I got, only one cable was able to transmit over 8mb/s. I'm no networking expert but the way I tested them was hooked up my laptop to my gigabit router and transferred a 4gb file to my unraid server. If results were slow, I'd flip the ends, reboot and repeat. Only one cable gave me ~50mb/s the rest were in the 2-8mb/s range. All signs point to the cables unless there may be something I may have overlooked. The funny thing is, I have a few older monoprice cat6 cables and they work great. However I don't want to take another gamble, so if anybody has any other suggestions for cables and where to get them, I would appreciate it.

I know its a dumb question, but is your laptop, switch and unRAID server all Gigabit?

Yes they all are. If they weren't I don't think I would be able to get ~50mb/s right? No settings were modified while I was testing the cables. The only thing that changed was that I switched ends and rebooted.

Have you tried switching the cables without a reboot in-between? Is it possible that a process/script is running at boot that is consuming network bandwidth/resources while you're testing?

 

Just a shot in the dark, but have you tried to transfer a file with the 50MB/s cable, swap for another cable, and transfer a different file from the same folder to the same remote folder and see the difference?

 

Seeing as I'm from the measure twice, cut once camp, I would also switch back to the 50MB/s cable and try a third file, just to confirm that the only variable changing is the cable itself...

 

(Also, make sure the files are suitably large, and similar - say 3 copies of the same iso file (with different names), for example).

 

I've never had a bad cable from monoprice, and I've bought significantly more then I'll ever need . It would be a real shame if this is a sign that their quality has dropped...

Have you tried switching the cables without a reboot in-between? Is it possible that a process/script is running at boot that is consuming network bandwidth/resources while you're testing?

 

Just a shot in the dark, but have you tried to transfer a file with the 50MB/s cable, swap for another cable, and transfer a different file from the same folder to the same remote folder and see the difference?

 

Seeing as I'm from the measure twice, cut once camp, I would also switch back to the 50MB/s cable and try a third file, just to confirm that the only variable changing is the cable itself...

 

(Also, make sure the files are suitably large, and similar - say 3 copies of the same iso file (with different names), for example).

 

I've never had a bad cable from monoprice, and I've bought significantly more then I'll ever need . It would be a real shame if this is a sign that their quality has dropped...

Yes I've tried switching cables without a reboot, and I would never get good speeds. It was only after a reboot, would I get decent speeds.

Too bad . Sounds like you've done everything you can, as far as confirming the issue is with the cables....

 

As far as another source, I've read good things in these forums about Superbiiz.com... You might want to give them a try...

Do you mean Megabits/s or MegaBytes/s?

 

Can you give us more about your tests? It sounds pretty vague at this point. Are you copying to an unRAID user share? Do you have a cache drive enabled for that share? What are the drive models/speeds/connections?

 

Booting shouldn't have anything to do with the negotiated connection or throughput, except perhaps randomly if things are borderline. In that case though you could get the same effect more quickly by disabling/enabling the interface. The only other thing a reboot would affect is that you might have processes starting at boot (or after being booted for awhile - depending on when it's slow) that are generating network traffic and affecting your tests.

 

Are you using the same switch/router ports? Is anything else happening at the router/switch? Have you tried different switches? How long are the cables? Which are they, 5e or 6? Are the cables bent, kinked, running around a refrigerator motor, etc.? How about the cable going to the server? Might it have a noisy connection but is being moved when you change the other cable?

 

Don't mean to nail you with 20 questions but these are the usual things that cross my mind.

If you are sure that the cables are bad, I would contact monoprice and request replacements.  I too have bought all my cables from monoprice and I've yet to see a bad one.

Do you mean Megabits/s or MegaBytes/s?

 

Can you give us more about your tests? It sounds pretty vague at this point. Are you copying to an unRAID user share? Do you have a cache drive enabled for that share? What are the drive models/speeds/connections?

 

Booting shouldn't have anything to do with the negotiated connection or throughput, except perhaps randomly if things are borderline. In that case though you could get the same effect more quickly by disabling/enabling the interface. The only other thing a reboot would affect is that you might have processes starting at boot (or after being booted for awhile - depending on when it's slow) that are generating network traffic and affecting your tests.

 

Are you using the same switch/router ports? Is anything else happening at the router/switch? Have you tried different switches? How long are the cables? Which are they, 5e or 6? Are the cables bent, kinked, running around a refrigerator motor, etc.? How about the cable going to the server? Might it have a noisy connection but is being moved when you change the other cable?

 

Don't mean to nail you with 20 questions but these are the usual things that cross my mind.

 

Megabytes per second.

 

I am taking a file on my win7 laptop hard drive, dragging and dropping directly into the user share. There is a cache drive, and from my understanding of Unraid, it is actually putting the data onto the cache drive and not one of the data drives. I am using the same EVERYTHING, which in my tests, consisted of a laptop, a gigabit router, and my unraid server. Cables are Cat6, all from monoprice. The only thing that changes is me unplugging a cable and switching it with a new one. Like I stated in my initial post, I have gotten good speeds with some older monoprice cables, but out of the current batch of cables, only one would give me the speed that I was used to. Cables are of varying length from 3ft-14ft and are not run around anything. I am literally sitting right next to the router as I'm plugging and unplugging the wires (same port each time).

 

Now I've heard of the occasional story of people receiving bad batches of their bulk cables (1000ft spools) so it wouldn't be to much of a stretch to imagine one bad spool getting cut up into a bunch of bad smaller cables.

Okay, good information. We're still missing a few things though. I mean, the copies are testing so many pieces at once that it's tough for me to blame anything specifically.

 

Do you know the speeds your notebook and unRAID are negotiating at the router? Does the router have LEDs that indicate 10/100 or 10/100/1000? Check it before and after any tests. Give it a few seconds then look for the Gb indicator on your router and/or the task manager pop-up in Windows. Try it a few times. Same results? If you see 100Mb connections then you know the problem and don't need to burn time copying files.

 

Cables can be bad. Spools can be bad. But it's tough for a short cable like those to be bad enough to make a 1Gb connection perform so poorly. More likely you'd see 100Mb negotiated connections and performance to match. But heck, maybe they only wired two pair? Reaching, I know.

 

The write speeds you report aren't that unusual. Not that it isn't telling us something, but it might not be. With my old slow parity drive I saw ~10MB/s to a user share over a Gb connection. My current cache drive gives me 55-60MB/s. But, if unRAID thinks the cache drive is full, depending on settings, it might be writing straight to the parity protected volume, slowing things down again. So, to be certain, test to \\servername\cache

 

If you consistently see Gb connections but performance is all over the place then we might want to dig more into error rates at both the server and notebook. Or, if you nail down some cables as always having the problem then maybe they are bad. It's just unusual and I'd be suspicious of something else failing or being borderline.

 

Just more thoughts, to use or ignore. Your choice.

You might want to use the software discussed in this post.  It tests bandwidth only - gets hard disk and some other factors out of the equation.  There are versions for Windows and Linux.  Recommend using the command line tool for Windows.

 

Iperf

You can try connecting the server and client back-to-back. Just to eliminate another variable.

You can try connecting the server and client back-to-back. Just to eliminate another variable.

 

Yep I would also suggest trying this,  once with your good cable and once with one of your duff ones.

 

 

 

You can try connecting the server and client back-to-back. Just to eliminate another variable.

 

Wouldn't that require a crossover cable as opposed to the regular patch cable that we all use as standard ethernet cables?

If you've got a Gigabit NIC in both computers, no need for a crossover.

If you've got a Gigabit NIC in both computers, no need for a crossover.

 

I'll take your word for it, I haven't tried to connect two computers directly in over 15 years.

If you've got a Gigabit NIC in both computers, no need for a crossover.

 

I'll take your word for it, I haven't tried to connect two computers directly in over 15 years.

 

Most modern gigabit devices automatically detect crossover or otherwise and sort it out automagically.  (Auto-MDIX)