Solved

How to resolve stalled inbound data transfers on sftp?

Posted on 2008-06-23
8
1,643 Views
Last Modified: 2010-04-21
A unique problem has presented itself, upon inbound transfers via sftp...once the first 6Kb of a file has been transferred the connection stalls. If left to run, eventually it will complete but only after stalling after every 6Kb of data transferred.

system:
2.6.22.14-72.fc6 running openssh-4.3p2-25.fc6

Data transfer tests with/without IPTABLES running produce same behavior (stalled transfers). Data transfer tests to identical OpenSSH versions on other boxes via Gigabit LAN do NOT produce this behavior (stalled transfers).

Has anyone seen this before or know of a solution?

Thanks,
0
Comment
Question by:mlsevResearch
[X]
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 4
  • 4
8 Comments
 
LVL 14

Expert Comment

by:agriesser
ID: 21850396
What do the error counters in `ifconfig` look like?
Such errors could be caused by duplex mismatches and/or bad cables. Can you try to manually set the port mode for testing purposes?

On Linux, you could use the old `mii-tool` or the new `ethtool` (whatever is installed on your system) to set the port mode manually.

Do you have a 100MBit switch (or configured switch port) available for testing purposes? Gigabit Ethernet over copper sometimes causes very strange issues if the cables are shielded and  I assume you're using shielded cables, right?
0
 

Author Comment

by:mlsevResearch
ID: 21850548
there are two boxes I'm receiving this error on so I'll paste sanitized output for both ifconfig & mii-tool:

box1:
ifconfig -> eth1
 RX packets:12696 errors:1221 dropped:0 overruns:0 frame:1221
 TX packets:4017 errors:0 dropped:0 overruns:0 carrier:0
mii-tool -> eth1
 eth1: negotiated 100baseTx-FD flow-control, link ok

box2:
ifconfig -> eth1
 RX packets:54276479 errors:7102653 dropped:0 overruns:0 frame:7102653
 TX packets:89790299 errors:0 dropped:0 overruns:0 carrier:0
mii-tool -> eth1
 eth1: negotiated 100baseTx-FD flow-control, link ok

all gigabit LAN traffic is for 192.168.0.0/27, connected to a LinkSys 8-port Gigabit switch

yes, I believe shielded cables are being used
0
 
LVL 14

Expert Comment

by:agriesser
ID: 21850614
Are these systems connected to the Gigabit switch? If so, mii-tool won't show the correct port settings as mii-tool is too old to know about GiE on copper ;)

OK, but we now know that there are no errors.
Try to switch the speed down to 100MBit/s or use unshielded cables for testing purposes (if available).
0
NFR key for Veeam Backup for Microsoft Office 365

Veeam is happy to provide a free NFR license (for 1 year, up to 10 users). This license allows for the non‑production use of Veeam Backup for Microsoft Office 365 in your home lab without any feature limitations.

 

Author Comment

by:mlsevResearch
ID: 21850687
the ifconfig output I posted shows both devices have RX errors...which is exactly what's happening.

not able to swap out cables for testing, will look into what port speeds I can set on the switch...

I would argue that it's the switch but with all devices connected to the same switch yet only two in the mix are having RX errors?? That doesn't make sense...
0
 
LVL 14

Accepted Solution

by:
agriesser earned 500 total points
ID: 21852935
You're right, sorry. I need to get myself some new glasses really soon...

If all machines connected to this switch would be of the same kind (same network card, same driver, etc.) you could definetly say that the cables are bad for the two machines which have the problem.

But in your case it could be everything (bad cables, buggy drivers, etc.).
Running GiE on the wire is a very delicate task where everything has to play together well.

In my office, f.ex., I needed to configure some ports of our Cisco Gigabit Switch Modules down to 100Mbit/s otherwise I wasn't able to get DHCP leases on them nor was I able to boot via PXE.

Besides that, using no real drivers but just the EEPROM of the NICs, loading software via PXE on Gigabit NICs is much slower than it is on 100MBit/s NICs using the already available cabling.
A month ago, we started to cable a new building and used unshielded CAT-7 cables which proofed the best performance I ever saw when using Gigabit Ethernet over copper. No problems in this building with PXE imaging, etc., it just simply works.
0
 

Author Comment

by:mlsevResearch
ID: 21855484
agriesser, thanks for the previous responses as well as the latest detailed response. I'm in the process of getting information together concerning the cabling utilized for the GiE set-up to see where we're at...

so hopefully this solution will be solved shortly and points awarded...
0
 

Author Closing Comment

by:mlsevResearch
ID: 31470316
power-cycle of switch cleared the problems, thanks for all the help
0
 
LVL 14

Expert Comment

by:agriesser
ID: 21859067
I recently RMA'ed a Linksys 8-port GiBit switch due to some ill effects like you experienced. You should definetly think about replacing this switch...
0

Featured Post

Industry Leaders: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Suggested Solutions

Title # Comments Views Activity
resolv.conf empty ubuntu server 14 3 75
Lame BIND 9.3 10 74
Linux SSH Error 9 161
Help needed with BIND9 DNS on Ubuntu. 22 105
I have seen several blogs and forum entries elsewhere state that because NTFS volumes do not support linux ownership or permissions, they cannot be used for anonymous ftp upload through the vsftpd program.   IT can be done and here's how to get i…
Note: for this to work properly you need to use a Cross-Over network cable. 1. Connect both servers S1 and S2 on the second network slots respectively. Note that you can use the 1st slots but usually these would be occupied by the Service Provide…
Finds all prime numbers in a range requested and places them in a public primes() array. I've demostrated a template size of 30 (2 * 3 * 5) but larger templates can be built such 210  (2 * 3 * 5 * 7) or 2310  (2 * 3 * 5 * 7 * 11). The larger templa…

726 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question