SMB+NFS keeps going unresponsive to clients

[scotch_tape]

OP deleted original question...

 

[dlavigne]

Anything related in /var/log/messages?

 

[anodos]

In addition to what Dru suggested, post the following:

• Hardware specs
• Type of network
• /usr/local/etc/smb4.conf file (enclosed in code brackets)
• interesting /var/log/samba4/log.smbd entries

 

[cyberjock]

You shouldn't be using 2_24. 2_10 is the default (and is what you should use). The other 2_xx that are above 2_10 are betas and such and shouldn't be used in production environments.

 

[cyberjock]

Sorry, I should clarify. The versions above 2_10 aren't true SMB versions in the normal sense. They were experimental versions of SMB that were used in Windows 8 betas. They shouldn't be used in production. I'm not even sure if Windows 8+ will use these weird experimental SMB versions, but I'm all about avoiding those weird situations when possible.

 

[anodos]

But Samba 3 came out 10 years ago? The windows 8 related Samba code didnt start until 3.5 in 2012.

SMB3 is not samba 3. http://en.wikipedia.org/wiki/Server_Message_Block

 

[anodos]

Thanks, I guess that is quite clear now. I wish the Freenas team would remove the newer versions as an option until such time as they are considered robust....

Defaults are there for a reason. From the outside it appears iX policy is to set safe defaults, but not create a padded room that prevents more advanced users from doing what they want. Setting server max protocol to SMB2_24 isn't going to cause your server to catch fire and lose your data. Samba may just be buggier than normal.

On a side-note, there are all sorts of ways a doggedly stupid person can mess up FreeNAS with minimal effort. It's just the nature of the beast. With great power comes great responsibility and all that jazz.

 

[anodos]

Not to be rude, but you've given enough info to narrow it down to either software config, hardware, or network config.

Need more info. Start with stuff requested in posts 2+3.

 

[cyberjock]

Then you have a failing disk. That would explain your server going unresponsive...

 

[cyberjock]

The dashboard doesn't report SMART errors. If you've setup emailing and such you should be getting an email every 30 minutes that your CUPS=10. I know I do. In fact I got a false high temp email just a few minutes ago from SMART from my 9.2.1.9 server.

 

[Ericloewe]

I'm becoming quite concerned. I have been religiously following the recommendations here, and now it appears that in order to discover a failed hard disk, I have to have performance problems and then discover failed disks via /var/log/messages ???

Have you confirmed that the e-mails are working?

 

[Ericloewe]

Yes, emails are working but apparently not with a failed disk?

What does smartctl -a /dev/whatever say about the disk in question?

 

[cyberjock]

Yes, emails are working but apparently not with a failed disk?

And you set the email box for the SMART settings page? Money says you didn't....

 

[BigDave]

My money says he doesn't have smart tests configured per recommended

 

[Jailer]

Dude, he's trying to help you narrow down the problem. You get sensitive and start spewing smart @$$ comments like that you can pretty much guarantee that you'll be on your own to troubleshoot your problem.

Don't forget, this is a community run forum. The ones posting are members of this forum and volunteers trying to help you sort out your issue with your FREE software. Just sayin.