15 replies to this topic

[brandobras]

Hello all,

suddenly since yesterday, i cannot reach my Hd51 with openpli 6 from my computer anymore.

i can see openwebif page if i type its ip number, but i can't transfer any files nor upload a channel list with dreambox edit.

if i try to reach it via filezilla i have a ECONNREFUSED error all the time and my hd51 is not shown anymore among my network devices

do you have any suggestions? both my pc and my hd51 are connected via ethernet to my router.

 

thank you

 

antonio

[WanWizard]

Filezilla = FTP? Network devices = Samba?

 

Is you box completely up to date? Did you reboot after the last update? Are you using 6.0-release, and not 6.0-rc?

[brandobras]

Hello,

I've updated the receiver and rebooted it and it works now!

 

thank you!!

[brandobras]

Hello,

I come back to the same topic as again the decoder is not visible in the network from my computer, but it's reachable via ip, FTP (filezilla), Telnet (Putty) and dreambox edit without any problem.

this times reboots (decoder, computer and router) didn't help.

it's not a real problem as far as i can reach it, but what happens is strange

 

best regards

[bril]

I have same problems when I want to look on my mutant1 for movies on the harddisk of my mutant2. With a few times trying, the harddisk in the other mutant is showing. Looks like a timeout problem in Samba. I not such a problem for me.

Edited by bril, 3 November 2017 - 11:40.

[WanWizard]

If this is about OpenPLi 6.0, you need to update until the box says there are no more updates, and then reboot the box. If not samba might not start, causing this issue.

[doglover]

I have the same on a HD2400.

Made a thread in the dutch section about this.:  https://forums.openp...mba-na-opstart/

 

Willy

[brandobras]

As for me it currently works, my HD51 just appeared in the list of devices connected to my network after few days and it's ok now.

 

best regards

[dog-man]

Since v6 release candidate and then the current v6 I have not had my VU showing in my network.

I can no longer use file explorer on my computer to se the files on the VU.

 

No issues FTP,ing files to and from it as I use DCC.

[WanWizard]

There have been at least 25 topics about this subject. 6.1. Will have a workaround for it.

[doglover]

As WW says, there more people who have the same problem.

It is not solved yet.  However there a couple of things you try.

 

Add to /etc/samba/smb.conf the following in the first section:

domain master = yes
preferred master = yes
local master = yes

The statement - domain master = yes - will force your receiver to be the master for samba.  This is fine when you have not defined somewhere else in your private network another samba master.  But if you have done that, you will know it, because you have added that statement to that linux machine.

 

After adding these statements restart samba on your receiver with the command:

/etc/init.d/samba.sh restart

You do this in a telnet window (DCC is OK)

 

Doing this, has improved the situation with several users.  In my own case the situation improved in so far that the error occurs now much less often.  And restarting samba cured the problem in each case.

 

It remains a problem, but at least you can correct it when it occurs.

 

You can also check the file /var/log/log.smdb from your receiver.  This should not contain any errors.

In case there are errors, try restarting samba again with the above statement.

 

Willy

Edited by doglover, 14 November 2017 - 13:53.

[WanWizard]

The statement - domain master = yes - will force your receiver to be the master for samba.  

 

Actually it is "local master" that does that. As the box is not a part of a domain, I don't think this line is needed.

 

I've added "local master" and "preferred master" to the default config for OpenPLi 6.1.

[littlesat]

I still have issues when a recording is made on the 'master' box and while the recording is still not ended watching the recording on a slave box...

The slave box stops playing the recording at the moment the recording was on when I started playing the recording...

 

E.g. when the recording takes 60 minutes in total... And I start playing it on the slave box at minute 30... the play stops at 30 minutes....

 

I can 'solve' this by showing the recording list in between (or via the console do a simple dir of the directory the recording is stored)...

This issue started as far I can see since we use a newer SAMBA... And I still did not find a configuration to 'improve' it so the end time of a recording movie is auto 'adjusted' on the slave box...

 

(on the master box there are of course no issues)

 

I hope you understand the issue as it is difficult to explain...

[WanWizard]

It is not a Samba issue, it is a known problem with modern CIFS kernel drivers. By default they cache stat info, so when Enigma asks for a file attribute update, the data from cache is returned, instead of asking the CFIS server for an update.

 

You can disable this mechanism on your mount, but that isn't very convinient, and impossible to "automate", similar to the differences in authtentication methods.

 

Could you try this on the client box:

echo 0 > /proc/fs/cifs/LookupCacheEnabled

and see if that improves things? And if not, try this one to

echo 0 > /proc/fs/cifs/LinuxExtensionsEnabled

and let me know if this helps?

 

I'm travelling, and have no means of testing it.

[littlesat]

If this helps.... Shouldn't we consider to disable this cache (which also has disadvantages)?

[WanWizard]

Yes. But I need to know first if it works, so if you can test this?
 
From a performance perspective, the best is to add "cache=loose" to the mount options.
 
The key issue is, as documented:

The default in kernels prior to 3.7 was "loose". As of kernel 3.7 the default is "strict".

Edited by WanWizard, 14 November 2017 - 16:45.