NAS lag when I go to a file share

[Jeggs101]

I have a small issue that is very annoying. At work we have a few Thecus, QNAP and now Synology units that are just inexpensive NAS units. They are local branch office caches of shared storage that backup to a central server.

When I go to these local NAS units, in office, and open the share \\SynologyNAS1 for example, there is a 10-20 second lag sometimes. Likewise, when I go to save a small Excel file at \\SynologyNAS1\myfiles I get the lag. Actual file transfers after the directory is loaded are around 100MB/s on gigabit and these Excel files are KBs not MBs.

I built a POC NAS based on NAS4Free last night (thanks to the STH main site article). Very similar setup but I used 2 6TB drives and an Intel S3700 400GB for L2ARC. Transfer speeds still around 100MB/s out of the box so that was nice. But the lag opening the share was still there, maybe even worse.

Does anyone else experience that kind of lag? Has anyone fixed it? Where do I start the searching?

 

[EffrafaxOfWug]

This is one of those "could be one of a million things" problems but a laundry list of things to check are:
a) your client, and the clients of other machines in those offices
b) DNS (forward and reverse)
c) if you're being old-skool, check netbios/nmbd
d) following on from b), check the domain name/FQDN component and ensure it doesn't clash with, amongst other things, someone running bonjour or some such nonsense on the local network. Try connecting by both the short name and FQDN.
e) Is the issue ever-present or does it stop once you've established a connection? Regularly flush your DNS cache and try reconnecting.
e) run a wireshark trace whilst you're trying to browse to the share. Is there traffic between you and the NAS the whole time during a stall? Are packets taking an aeon to return? Is stuff flinghoodling off in entirely the wrong direction?
f) check non-CIFS network performance. Is HTTP, ICMP, SSH traffic to the NAS similarly sluggish?

 

[NetWise]

G) does the NAS, drives, or services go to slee and have to spin up or launch as each connection is made?

H) are they workgroup or domain? Could the issue be doing an authentication lookup or trying to use some pass through credential then timing out, using a second credential and working?

 

[Jeggs101]

The network in each office is <20 machines so it isn't a load issue. No vlans no hubs, all into the same switch. I actually tried this even at home using the default "workgroup" and by using the IP address of the machine instead of the name. Credentials saved locally.

Does anyone else have this issue by chance?

 

[mrkrad]

Do you have IPV6 enabled on those pc's? It can cause severe lag with improper routers/switches/setup

 

[markarr]

Careful disabling IPv6 on any windows server. Several of microsoft services are using IPv6 for communication between them, one example if you disable IPv6 on an exchange server it stops working even if you never setup any IPv6 on your network. Good ol Microsoft.

 

[mrkrad]

Yeah I assume windows 7/8 clients - Don't disable ipv6 with any microsoft services (AD/DNS/exchange etc). But for improperly setup networks ipv6 on clients will cause instability/lag on the network stack!

 

[HellDiverUK]

I did have a similar issue at work, once or twice in 6 years. Restarting DNS on the domain controller fixed it.

 

[Patrick]

One other thing to check via the CLI - is the NIC working. I had an issue on that due to a slightly overheating NIC.

 

[mattr]

G) does the NAS, drives, or services go to slee and have to spin up or launch as each connection is made?

I would second this. I have a DS214+ that I'm playing with now and even though all the hibernation settings are disabled and the HDD firmware has power management disabled the drives still spin-down. The same drives never spin-down in my file server or my Thecus Debian FS. I actually don't use the DS214+ because as intended because of the spin-down.