FreeNAS 11.1-U6 Reset all of my container configs on RancherOS

[Spencer Skinner]

Has anyone else had there containers basically reset back to first install by the update, I've got no idea why its happened as all my configs are stores in persistent storage outside the RancherOS VM. Im now all of a sudden having issues with Rancher being unable to see inside certain directories that it previously could access before the update.

Any ideas on this?

 

[KrisBee]

What protocol did you use to share pool data with your docker containers? NFS or CIFS? Could this be down to the SAMBA changes in the update?

https://www.ixsystems.com/blog/library/freenas-11-1-u6/

 

[Spencer Skinner]

What protocol did you use to share pool data with your docker containers? NFS or CIFS? Could this be down to the SAMBA changes in the update?

https://www.ixsystems.com/blog/library/freenas-11-1-u6/

I used CIFS, so id imagine that the SAMBA changes have effected this, could it be the SMBv1 being disabled potentially? I feel like this must be the case as my CIFS shares no longer appear as mounted when I run sudo df -h

Code:

mounts:
- - //192.168.2.10/AlphaData_GlobalShare
  - /media/AlphaData
  - cifs
  - username=spencer,password=xxxxxx,iocharset=utf8,_netdev
- - //192.168.2.10/AlphaMedia_GlobalShare
  - /media/AlphaMedia
  - cifs
  - username=spencer,password=xxxxxx,iocharset=utf8,_netdev
- - //192.168.2.10/VMRoot_GlobalShare
  - /media/VMRoot
  - cifs
  - username=spencer,password=xxxxxx,iocharset=utf8,_netdev

That is the contents of my mounts.yaml file which i then merge with the config using sudo ros config merge -i mount.yaml

Could be worth enabling SMBv1 to test it? What are your thoughts?

 

[KrisBee]

You could either add a "vers=2.1" option to all your CIFS mount in the yaml file, or start by adding a min protocol statement to the auxillary params of your FreeNAS SMB service settings. There's been several thread about it on the forum.

 

[Spencer Skinner]

You could either add a "vers=2.1" option to all your CIFS mount in the yaml file, or start by adding a min protocol statement to the auxillary params of your FreeNAS SMB service settings. There's been several thread about it on the forum.

Have multiple others been having issue like mine then, I wasnt aware?

Whats the structure of a min protocol statement is it server_min_protocol=NT1

Thanks

 

[Spencer Skinner]

You could either add a "vers=2.1" option to all your CIFS mount in the yaml file, or start by adding a min protocol statement to the auxillary params of your FreeNAS SMB service settings. There's been several thread about it on the forum.

Have multiple others been having issue like mine then, I wasnt aware?

Whats the structure of a min protocol statement is it server_min_protocol=NT1

Thanks

Right so ive tried dropping the min protocol down to V1 like I showed in my previous reply, I also added the version statement as you showed. It still hasnt started showing the mounts when I do sudo df -h

Code:

mounts:
- - //192.168.2.10/AlphaData_GlobalShare
  - /media/AlphaData
  - cifs
  - vers=2.1
  - username=spencer,password=xxxx,iocharset=utf8,_netdev
- - //192.168.2.10/AlphaMedia_GlobalShare
  - /media/AlphaMedia
  - cifs
  - vers=2.1
  - username=spencer,password=xxxx,iocharset=utf8,_netdev
- - //192.168.2.10/VMRoot_GlobalShare
  - /media/VMRoot
  - cifs
  - vers=2.1
  - username=spencer,password=xxxx,iocharset=utf8,_netdev

 

[KrisBee]

Don't know about other people having the same problem as yourself. I'm not actively using rancheros myself, you'll have to resort to logs and the usual client/server side commands to check whats going on. Become root on rancheros with sudo -i and have a poke around in the rancheros logs. Can you manually mount a CIFS share, etc., as root on rancheros? Server side smbstatus and smbclient / smbtree and SAMBA logs might offer some clues.

 

[Spencer Skinner]

Don't know about other people having the same problem as yourself. I'm not actively using rancheros myself, you'll have to resort to logs and the usual client/server side commands to check whats going on. Become root on rancheros with sudo -i and have a poke around in the rancheros logs. Can you manually mount a CIFS share, etc., as root on rancheros? Server side smbstatus and smbclient / smbtree and SAMBA logs might offer some clues.

Ive tried checking the SMB status within rancher and none of those commands seem to work, it just throws command not found. Ive also tried manually mounting the share with sudo mount -t cifs -o username=spencer,password=xxxxx//192.168.2.10/AlphaMedia_GlobalShare /mnt and that throws and operation not supported error.

 

[KrisBee]

SMB commands were for server side, i.e FreeNAS. Does just mount show any CIF mounts on rancheros? What about logs? tail/grep /var/log/messages on rancheros and /var/log/samba4/log.smbd on FreeNAS

 

[Spencer Skinner]

SMB commands were for server side, i.e FreeNAS. Does just mount show any CIF mounts on rancheros? What about logs? tail/grep /var/log/messages on rancheros and /var/log/samba4/log.smbd on FreeNAS

The RancherOS logs dont show anything about SMB but here is what i got

Code:

[rancher@Rancher ~]$ sudo tail /var/log/messages
Oct  1 17:05:17 Rancher kernel: [  748.619996] device vethr6e401a2059 entered promiscuous mode
Oct  1 17:05:17 Rancher kernel: [  748.620019] docker0: port 4(vethr6e401a2059) entered blocking state
Oct  1 17:05:17 Rancher kernel: [  748.620019] docker0: port 4(vethr6e401a2059) entered forwarding state
Oct  1 17:05:20 Rancher ntpd[1]: Listen normally on 19 vethr6e401a2059 [fe80::482d:6cff:fe9f:4c6%13]:123
Oct  1 19:36:00 Rancher kernel: [ 9792.331091] clocksource: timekeeping watchdog on CPU1: Marking clocksource 'tsc' as unstable because the skew is too large:
Oct  1 19:36:00 Rancher kernel: [ 9792.331093] clocksource:					   'refined-jiffies' wd_now: 100240df8 wd_last: 100240d78 mask: ffffffff
Oct  1 19:36:00 Rancher kernel: [ 9792.331094] clocksource:					   'tsc' cs_now: 44eac5c080e72 cs_last: 44eabd7d0b964 mask: ffffffffffffffff
Oct  1 19:36:00 Rancher kernel: [ 9792.332043] clocksource: Switched to clocksource refined-jiffies
Oct  1 20:06:25 Rancher kernel: [11623.566058] serial8250: too much work for irq4
Oct  1 20:06:25 Rancher kernel: [11623.597817] serial8250: too much work for irq4

Where as the FreeNAS logs show that my min version parameter doesnt work, not sure what i should change it to

Code:

root@AlphaServer:~ # tail /var/log/samba4/log.smbd
[2018/10/01 18:10:17.432821,  0] ../lib/param/loadparm.c:1811(lpcfg_do_global_parameter)
  Ignoring unknown parameter "server_min_protocol"
[2018/10/01 19:39:46.002410,  0] ../source3/lib/util_sock.c:876(matchname)
  matchname: host name/name mismatch: 192.168.2.111 != (NULL)
[2018/10/01 19:39:46.002459,  0] ../source3/lib/util_sock.c:1055(get_remote_hostname)
  matchname failed on 192.168.2.111
[2018/10/01 19:39:46.006142,  0] ../lib/param/loadparm.c:784(lpcfg_map_parameter)
  Unknown parameter encountered: "server_min_protocol"
[2018/10/01 19:39:46.006171,  0] ../lib/param/loadparm.c:1811(lpcfg_do_global_parameter)
  Ignoring unknown parameter "server_min_protocol"

Mount on RancherOS also shows nothing about any mounted CIFS shares, only mounts inside the RancherOS

Thanks

 

[KrisBee]

You need to look through logs for past entries and/or watch logs as new entries are made. The tail command can be used like this:

EXAMPLES
To display the last 500 lines of the file foo:

$ tail -n 500 foo

Keep /var/log/messages open, displaying to the standard output anything
appended to the file:

$ tail -f /var/log/messages

A simple less /var/log/messages allows use of the up/down arrow to move through text files.

Search the forums for the correct syntax of the auxillary param you need to add. The last few entries of this thread suggests other have had similar(?) problems with CIFS mounts in rancheros: https://forums.freenas.org/index.ph...-rancheros-docker-rancherui-plex.59963/page-6

Personally I wouldn't use rancheros for small scale docker use, I'd prefer a debian/ubuntu based VM combined with docker and portainer for container management.

 

[Spencer Skinner]

You need to look through logs for past entries and/or watch logs as new entries are made. The tail command can be used like this:



A simple less /var/log/messages allows use of the up/down arrow to move through text files.

Search the forums for the correct syntax of the auxillary param you need to add. The last few entries of this thread suggests other have had similar(?) problems with CIFS mounts in rancheros: https://forums.freenas.org/index.ph...-rancheros-docker-rancherui-plex.59963/page-6

Personally I wouldn't use rancheros for small scale docker use, I'd prefer a debian/ubuntu based VM combined with docker and portainer for container management.

I cant find anything on the SMB logs so far showing nothing about rancher. Same goes for the logs on RancherOS, they dont seem to show anything either. I found the correct syntax and tried that but that didnt work either. Ive taken a look in that post too, popped a question there.

I feel like you might be right on the from of Ubuntu and portainer to be honest, as I will have the free will to do updates to portainer and docker when they are released which i dont have right now with rancherOS.

How simple would that be to set up?

Thanks

 

[KrisBee]

It's odd that you've not found any indication of why the CIFS Mounts are failing om rancheros. There are various logs under /var/log and /var/log/boot you could check.

How simple would that be to set up?

Simple is relative to your skills etc. But in FreeNAS 11.1-U6 if you stick with Ubuntu 16 LTS, creating the VM should be straightforward. A bit of googling will show you how to add the latest docker, and docker compose it you want, to Ubuntu. Then follow portainer's website to add their container management app. Of course this means recreating your containers and linking them to your pool data. So if you were storing any config data etc. inside your original containers that would be lost.

 

[Spencer Skinner]

It's odd that you've not found any indication of why the CIFS Mounts are failing om rancheros. There are various logs under /var/log and /var/log/boot you could check.



Simple is relative to your skills etc. But in FreeNAS 11.1-U6 if you stick with Ubuntu 16 LTS, creating the VM should be straightforward. A bit of googling will show you how to add the latest docker, and docker compose it you want, to Ubuntu. Then follow portainer's website to add their container management app. Of course this means recreating your containers and linking them to your pool data. So if you were storing any config data etc. inside your original containers that would be lost.

Yeh im going to see what comes back about my config file, see if im the one causing the error ahaha.

As for Ubuntu, im gonna give that a try in the meantime, would you recommend going for ubuntu server or desktop for this, I've assumed server so far, had a few issue with installing desktop, server seems to be more successful so far. Yeh as it stands right now im basically going to have to start fresh with all my containers anyway so I dont mind a full swap over at this point

 

[KrisBee]

Ubuntu server would be my choice, you don't need a desktop gui and portainer provides a webUI for container management.

 

[Spencer Skinner]

Ubuntu server would be my choice, you don't need a desktop gui and portainer provides a webUI for container management.

Ive managed to solve the issue with Rancher now, it turns out you need to add vers=3.0 as a parameter to the mount. I realised this when i was doing the mounts for my ubuntu VM, that said I'm likely to migrate over to the portainer setup you've suggested as it is far easier to maintain and will receive more regular updates than my rancher setup due to the nature of how it was deployed. Thanks allot for your help

 

[KrisBee]

Thanks for reporting back. Some what late, but I remembered there are some ideas about using portainer in the Forum's resources section.