PDA

View Full Version : Persistent network mount possible?



edwin2006
2017-07-22, 08:38
Using lms and local music is on windows share.
Mounting is successful and I can drill down to the right folder.
But after some time listening to internet radio I'm unable to play local music files.
If I check folder in setup I'm unable to drill down the mounting point.
Reboot pi does not help. Only thing is to first intentionally use the wrong password to mount, giving an error, and then remounting with correct password.
After that it works some time, I guess until I'm using internet radio for a while and therefore not using the mounting point.
I see an option field behind the mounting point but not having a clue what options are available. Perhaps one to persistently mount the cifs share?

castalla
2017-07-22, 08:49
Using lms and local music is on windows share.
Mounting is successful and I can drill down to the right folder.
But after some time listening to internet radio I'm unable to play local music files.
If I check folder in setup I'm unable to drill down the mounting point.
Reboot pi does not help. Only thing is to first intentionally use the wrong password to mount, giving an error, and then remounting with correct password.
After that it works some time, I guess until I'm using internet radio for a while and therefore not using the mounting point.
I see an option field behind the mounting point but not having a clue what options are available. Perhaps one to persistently mount the cifs share?

Given all the issues you are having, I suggest that the core problem lies in how your network is setup and run via the Fritzbox and the switch. No idea how to fix it though!

edwin2006
2017-07-22, 09:04
Perhaps that's something. Didn't reboot the router perhaps that's the problem even if I'm only using the fritz box as gateway to internet. So only one Lan port in use. Banging my head to the wall about this 😠
Will try reboot fritz box.

d6jg
2017-07-22, 11:21
Is the Windows machine configured not to go into sleep mode or power down the disks?

edwin2006
2017-07-22, 12:35
Absolutely, it's running 24/7

drmatt
2017-07-22, 23:09
Two things. What distro on the pi and how did you setup the cifs mount? Typically a cifs mount involves you storing the password in the fstab. If you do this it will always reconnect after any network glitch. If you don't do this, it will not reconnect after a reboot, but should still maintain the mount until you explicitly unmount it.


Transcoded from Matt's brain by Tapatalk

edwin2006
2017-07-23, 01:37
I'm using PiCorePlayer v3.20

d6jg
2017-07-23, 03:32
Absolutely, it's running 24/7

My money is still on the Windows box being the issue. I have never experienced such an issue with pCP mounting a Linux based share on a NAS. The only caveat being that the NAS needs to be fully up and running before you boot the Pi.

kidstypike
2017-07-23, 03:59
My money is still on the Windows box being the issue. I have never experienced such an issue with pCP mounting a Linux based share on a NAS. The only caveat being that the NAS needs to be fully up and running before you boot the Pi.

I agree. My music files are on a WD myCloud (that's it below next to my LMS/piCorePlayer sever). It's set to sleep after a period of non activity, it wakes no problem when there's a demand. It's no trouble whatsoever, it just works.

23204

23205

paul-
2017-07-23, 06:43
The only caveat being that the NAS needs to be fully up and running before you boot the Pi.

With 3.21, there is now a retry loop on the network mount, to handle reboots from a power outage. I think it will try for about 2 minutes. Nas boxes normally startup pretty quick.

edwin2006
2017-07-23, 08:44
My money is still on the Windows box being the issue. I have never experienced such an issue with pCP mounting a Linux based share on a NAS. The only caveat being that the NAS needs to be fully up and running before you boot the Pi.
Well did some testing.
1. i disabled all non-logitech plug-ins, so also no spotty => don't think thats helping
2. Copied some music to a LaCie network 2 single disk nas. Used same name & credentials for the share. Disk spin down after 5 minutes. => works like it should
3. Enabled spotty => everything works. if i use internet radio or spotty for a while i can switch back to local files.
4. Created a share on a different W7-Pro box with same share name and credentials => listening to internet radio very quickly the local files cannot be used.
5. Back to share on Lacie => everything works.
6. Using share from 4 but just listening to local files => no problem, keeps working.

So my conclusion (and correct me if im wrong, please) is that somehow the mounting point of LMS does not work well when mounted to a windows box. I tried two different windows computers.
Also the strange problems with spotty being not able to search for a number and then play it are solved when using the LaCie.
The easiest way would be to move music etc to a NAS but i would like to keep a single point for storage to make backup's etc easy. So keeping the music on the windows box.

Now the golden question. Can this be solved and how?

drmatt
2017-07-23, 08:51
Simple answer: mount the Nas to the windows box and you have a single point of storage.. :)

No, there is no fundamental flaw in the cifs client, there is a problem with Auth or the SMB service on your Windows box.


Transcoded from Matt's brain by Tapatalk

edwin2006
2017-07-23, 09:07
It takes about 15 minutes before is looses the mounting point and then LMS is going down. Only reboot pi solves the problem.
That is when I'm not playing local files.

epoch1970
2017-07-23, 09:59
This looks like your issue. https://support.microsoft.com/en-us/help/297684/mapped-drive-connection-to-network-share-may-be-lost

The server drops inactive connection on its side (rather than waiting for a client disconnect), probably so that it can spin-down drives or sleep even if clients were not disconnected properly.
So instead of rebooting, dismounting/remounting on the client should suffice.

edwin2006
2017-07-23, 12:09
@epoch1970.
Thanks, i searched but only find a solution for the client side.
I totally switched autoconnect off with net config server /autodisconnect:-1 and let's see what happens.

Still i think something is working not as expected at the LMS/PcP side. If it just would reconnect then there should be no issue but somehow linux and windows misunderstand eachother. From another windows box there is no problem with this share.

drmatt
2017-07-23, 12:18
But others can mount windows shares without drops.. I can. It's just a cifs mount in the Linux kernel, this is not rocket science.


Transcoded from Matt's brain by Tapatalk

castalla
2017-07-23, 12:36
@epoch1970.
From another windows box there is no problem with this share.

... which suggests the problem is with the other win box in how the shares are configured in windows.

d6jg
2017-07-23, 12:51
Are both the Pi and the Windows box wired? If one or both is wireless this may be the cause.

edwin2006
2017-07-23, 14:00
Yes, both are wired.
I don't want to get in to a linux / windows bashing so of course linux is superior and windows sucks but never the less it is odd that the temporally closing of a share causes LMS / PcP to get this up-said.
Would be nice if this solved itself when the share is not ready as expected.

atrocity
2017-07-23, 15:25
Yes, both are wired.
I don't want to get in to a linux / windows bashing so of course linux is superior and windows sucks but never the less it is odd that the temporally closing of a share causes LMS / PcP to get this up-said.
Would be nice if this solved itself when the share is not ready as expected.

I used to have a Windows Vista box hosting a shared drive. It would always work for a while, sometimes for days, but it would always eventually stop sharing. I found a registry hack and that fixed the problem. But then my computer died and I had to buy another one that had Windows 7. Hosting shared drives had the same issue and in that case the hack didn't work. That was an important nudge to get me to use Linux more.

More recently--and I wonder if this is relevant to your issue--Ubuntu (at least) Linux had a bug where cifs mounts in fstab would become unavailable pretty quickly and reliably. Weirdly, the workaround was to add vers=[version number] as a mount option. As I adopted the workaround, I am not aware if the actual bug has been fixed yet, nor whether piCorePlayer and Ubuntu share any code.

https://askubuntu.com/questions/915549/16-04-cifs-host-is-down-but-they-are-not

edwin2006
2017-07-23, 23:51
No idea. At first when i saw noticed SMB1 in the article i thought "that must have to do with shutting down SMB1 on windows due to the wannacry outbreak wich forced microsoft to disable SMB1 by default" but then i read "To access a share that is "down" takes about 2-3min of randomly clicking different mounts" and now i'don't know.
Somewhere in the article i also noticed "https://wiki.ubuntu.com/MountWindowsSharesPermanently" which is bit over my head and pretty old, especially since i don't know how the PcP setup internally is.

I think someone with inside knowledge of PcP and LMS need to look into this. As said it is bit strange that LMS is getting this much up said over a non accessible share since i'm only using it for the music files and playlist. All other paths are local on the pi (cache, prefs, plugin etc.)

castalla
2017-07-24, 00:11
No idea. At first when i saw noticed SMB1 in the article i thought "that must have to do with shutting down SMB1 on windows due to the wannacry outbreak wich forced microsoft to disable SMB1 by default" but then i read "To access a share that is "down" takes about 2-3min of randomly clicking different mounts" and now i'don't know.
Somewhere in the article i also noticed "https://wiki.ubuntu.com/MountWindowsSharesPermanently" which is bit over my head and pretty old, especially since i don't know how the PcP setup internally is.

I think someone with inside knowledge of PcP and LMS need to look into this. As said it is bit strange that LMS is getting this much up said over a non accessible share since i'm only using it for the music files and playlist. All other paths are local on the pi (cache, prefs, plugin etc.)

Not sure about others, but I'm very confused by how your setup works!

You have a pi with picore

You have a La Cie drive - shares work

You have a win7 box - shares fail

You have a 2nd win7 box - shares work

Is this correct?

slartibartfast
2017-07-24, 00:29
This is probably a stupid question but I assume there must be a good reason why you don't run LMS on the Windows box.

Sent from my SM-G900F using Tapatalk

edwin2006
2017-07-24, 00:36
Just found out that while the share mount survived 12 hours with the autoreconnect-1 etc change the netwerkshare mount did not get reconnected when i rebooted the pi. So still there is something not allright.
Now for your question.

I've got an older WXP system that is only being used as fileshare/domotica/LMS/music ripping station "server" bit i disabled LMS at this moment via wired network. Will be replaced, but not at this moment.
A Pi3 with PiCorePlayer 3.20 on which LMS is enabled and Squeezeplayer is disabled, via wired network (no wifi)
one LaCie drive that is in use for backup of XP station. Temporarily is created a share with some music files for testing networkshare mount.
one W7 system that is in use as workstation, wired network. Normally no shares but for testing this problem i created a share just as on the Lacie.
At the moment of testing all connected via a single "dumb" netgear GS116 (10/100/1000 16p switch)
All on static IP outside DHCP range.

The linux lacie box => no problem, share stays connected even when rebooting pi.
The XP and W7 box => disconnect problem after a while not using the share but this seems to be solved with the autoreconnect-1 etc modification. Does not survive with pi reboot

So when the pi has a reboot i have to go to the lms tab, change ip adres for the share mount to the ip adres of the lacie (with temp share), and click on the "Set NET Mount". return to main page, and change ip adres to XP box and click on the "Set NET Mount". Then the network mount works and keeps on working (due to the autoreconnect change) until the pi get's rebooted. If i do it different i get the notification nothing changed and the mounting point is not refreshed.

Somehow, somewhere the mounting process during pi boot is different compared to the manually "Set NET Mount" which causes the mount to fail after reboot.

edwin2006
2017-07-24, 00:37
This is probably a stupid question but I assume there must be a good reason why you don't run LMS on the Windows box.

Sent from my SM-G900F using Tapatalk
Yep, XP box. No Spotty support.

d6jg
2017-07-24, 00:43
Just found out that while the share mount survived 12 hours with the autoreconnect-1 etc change the netwerkshare mount did not get reconnected when i rebooted the pi. So still there is something not allright.
Now for your question.

I've got an older WXP system that is only being used as fileshare/domotica/LMS/music ripping station "server" bit i disabled LMS at this moment via wired network. Will be replaced, but not at this moment.
A Pi3 with PiCorePlayer 3.20 on which LMS is enabled and Squeezeplayer is disabled, via wired network (no wifi)
one LaCie drive that is in use for backup of XP station. Temporarily is created a share with some music files for testing networkshare mount.
one W7 system that is in use as workstation, wired network. Normally no shares but for testing this problem i created a share just as on the Lacie.
At the moment of testing all connected via a single "dumb" netgear GS116 (10/100/1000 16p switch)

The linux lacie box => no problem, share stays connected even when rebooting pi.
The XP and W7 box => disconnect problem after a while not using the share but this seems to be solved with the autoreconnect-1 etc modification. Does not survive with pi reboot

So when the pi has a reboot i have to go to the lms tab, change ip adres for the share mount to the ip adres of the lacie (with temp share), and click on the "Set NET Mount". return to main page, and change ip adres to XP box and click on the "Set NET Mount". Then the network mount works and keeps on working (due to the autoreconnect change) until the pi get's rebooted. If i do it different i get the notification nothing changed and the mounting point is not refreshed.

Somehow, somewhere the mounting process during pi boot is different compared to the manually "Set NET Mount" which causes the mount to fail after reboot.

You need to do aconfig backup after the mount specifically so that it survives a reboot

edwin2006
2017-07-24, 00:50
Eh, when i use the web interface for set NET Mount it is doing a config backup or am i wrong?
After reboot pi i see the correct IP adress (XP Box) and credentials in the PcP LMS Section but am unable to drill down via (IPadres:9000/settings) LMS Server setting -> Media Folder, so it has not really been doing a fresh mount. Only via workaround with actually changing IP will help.

epoch1970
2017-07-24, 01:05
Just in case. 12 hrs sounds like a dhcp lease length. For sure the server win machine should use a fixed IP and infinite length lease, perhaps the client needs this too?

edwin2006
2017-07-24, 01:25
Shoot, forgot to mention. All on static IP outside DHCP range. Will add to previous post.

d6jg
2017-07-24, 02:34
Eh, when i use the web interface for set NET Mount it is doing a config backup or am i wrong?
After reboot pi i see the correct IP adress (XP Box) and credentials in the PcP LMS Section but am unable to drill down via (IPadres:9000/settings) LMS Server setting -> Media Folder, so it has not really been doing a fresh mount. Only via workaround with actually changing IP will help.

So you mean the mount point persists but the Pi can't see the XP box.
I'd say again, thats the Windows box not the Pi.
XP has been dead for a few years now and as such I think you are flogging a dead horse so to speak.
If the Lacie NAS works repurpose it as a central store not a backup and use it (or get another NAS for LMS).
This is one of those situations where it really isn't worth anyone's effort to find a solution because the XP box is obsolete.
Had you mentioned it was XP at the outset it would have saved a lot of time but you didn't tell us that until post 24!!
This started off because Spotty won't work on the XP box didn't it? Do you expect Michael to waste time trying to make Spotty run on it ?

edwin2006
2017-07-24, 02:45
See after => in red

So you mean the mount point persists but the Pi can't see the XP box. => don't know, unable to drill down the mounting point in settings LMS
I'd say again, thats the Windows box not the Pi. => same problem at W7 and XP box, not so likely that windows is the culprit
XP has been dead for a few years now and as such I think you are flogging a dead horse so to speak. => i know, that's why going pi route
If the Lacie NAS works repurpose it as a central store not a backup and use it (or get another NAS for LMS). => XP box is going to be replaced but i guess this will have the same mounting problem since W7 does.
This is one of those situations where it really isn't worth anyone's effort to find a solution because the XP box is obsolete. => not looking for XP solution, just mounting solution
Had you mentioned it was XP at the outset it would have saved a lot of time but you didn't tell us that until post 24!! => if i did not, sorry for that
This started off because Spotty won't work on the XP box didn't it? Do you expect Michael to waste time trying to make Spotty run on it ? => No, that's why i went the pi route for LMS and then stumbled upon the mounting issue

castalla
2017-07-24, 04:37
Just found out that while the share mount survived 12 hours with the autoreconnect-1

How are you specifying the autoreconnect ?

edwin2006
2017-07-24, 04:40
Allright, took another box and installed windows server 2016 essentials. Created a share (with the same name as before) and a user (same credentials).
Goal was to exclude windows version problem. Thinking if i took a box that is ment to share files it would be the best bet. Let's see if this one also got disconnected over time since a few hours after my post about the success of the autoconnect=-1 option it seems to got lost again.

W7 box is perfectly able to access share, PcP unfortunately not. Tried the vers=2.0 AND vers=3.0 both without luck.
So i'm clueless at this moment. I still think there is something off with automatically reconnecting the mounting point when a windows box is providing the share.

drmatt
2017-07-24, 04:44
Windows clients always transparently reconnect on bad links by the way, they hide the failure from view. I don't have any further experience at home, my servers are all Linux and I only have windows as client. In the past I've not had the issue you mention however.

Random comment but is there significant clock drift on your machines?


Transcoded from Matt's brain by Tapatalk

edwin2006
2017-07-24, 05:07
Windows clients always transparently reconnect on bad links by the way, they hide the failure from view. I don't have any further experience at home, my servers are all Linux and I only have windows as client. In the past I've not had the issue you mention however.

Random comment but is there significant clock drift on your machines?


Transcoded from Matt's brain by Tapatalk
Good question but all via NTP. Just checked all (W7/XP/W2k16) in sync.
btw. Checked log files PcP after failed mount to w2k16 share but found nothing?
Also tried accessing the w2k16 share from a ubuntu 16.04 system and that is no problem. I can see and play the files.

castalla
2017-07-24, 06:58
Good question but all via NTP. Just checked all (W7/XP/W2k16) in sync.
btw. Checked log files PcP after failed mount to w2k16 share but found nothing?
Also tried accessing the w2k16 share from a ubuntu 16.04 system and that is no problem. I can see and play the files.

I dragged out my Picoreplayer ... set share on Win7. Reboots of picore don't affect the mounted share.

edwin2006
2017-07-24, 07:02
I dragged out my Picoreplayer ... set share on Win7. Reboots of picore don't affect the mounted share.
That's correct (i think). If it's not connected anymore a reboot will not help you out.

castalla
2017-07-24, 07:23
That's correct (i think). If it's not connected anymore a reboot will not help you out.

I'm now totally lost. What do you mean?

Isn't your problem that the share is lost within minutes (not on my system), that a reboot of pcp kills the share (not on my system) ?

[Having resurrected picoreplayer, I find it's a really confusing interface to mess around with compared to max2play]

edwin2006
2017-07-24, 07:43
No problem.
When I use a share on a W7 box and mount that share from PcP everything is ok at first.
After a while the W7 box closes the share (autodisconnect function) and that is where PcP comes into trouble. When you try to drill down mediafolders it just stops at the mounting point and won't let you go deeper.
At this point PcP/LMS is totally unresponsive. I think it should (an perhaps does but fails) try to reconnect but is does not succeed.
Switching off the autodisconnect seemed to help but I'm now unsure because it happened, a very long time later, again. If you keep on playing local files everything is ok as far as I know.
btw the autodisconnect takes time, i think something like 30 minutes and only happens when you're playing internet stuff (radio, spotty). When you changed the W7 system per MS article it takes a lot longer, if at all. It is now running almost 3 hours and the mounting point still works (XP share since there are all my music files).

castalla
2017-07-24, 07:47
No problem.
When I use a share on a W7 box and mount that share from PcP everything is ok at first.
After a while the W7 box closes the share (autodisconnect function) and that is where PcP comes into trouble. When you try to drill down mediafolders it just stops at the mounting point and won't let you go deeper.
At this point PcP/LMS is totally unresponsive. I think it should (an perhaps does but fails) try to reconnect but is does not succeed.
Switching off the autodisconnect seemed to help but I'm now unsure because it happened, a very long time later, again. If you keep on playing local files everything is ok as far as I know.
btw the autodisconnect takes time, i think something like 30 minutes and only happens when you're playing internet stuff (radio, spotty). When you changed the W7 system per MS article it takes a lot longer, if at all. It is now running almost 3 hours and the mounting point still works (XP share since there are all my music files).

Okay - I'll set it to play radio for a few hours, then check the file mount ...

edwin2006
2017-07-24, 07:58
thnx.

castalla
2017-07-24, 09:31
thnx.

So, played radio stream for about 30 minutes - switched to music, and no issues playing from the shared folder on Win7

atrocity
2017-07-24, 09:50
That's correct (i think). If it's not connected anymore a reboot will not help you out.

Next time you have the problem, can you SSH into pCP and issue a
sudo mount -a command?

edwin2006
2017-07-24, 10:23
Absolutely. What would be the result, I assume it is supposed to do a reconnect?

castalla
2017-07-24, 10:40
Absolutely. What would be the result, I assume it is supposed to do a reconnect?

It does a remount ... but doesn't fix the underlying problem

d6jg
2017-07-24, 10:48
I'd suggest you run

df -h

Then

mount -a

The former will display all of the top level folders including anything that's mounted from elsewhere

The latter tries to mount whatever is specified in fstab

The output will show us a) if the Pi thinks the folder is already mounted and b) display an error message if the remount fails

edwin2006
2017-07-24, 11:09
Will do. At the moment the mount is holding more than 6 hours.

d6jg
2017-07-24, 12:49
Edwin
I have just spotted your thread about SB Radio disappearing from LMS in SB Radio area.
This and all the other things you are experiencing says network issues.
I know I asked before and you replied to say everything is wired but I will ask again - are the Windows XP box and LMS both wired directly to the same switch? Is there anything in between? Anything ?

edwin2006
2017-07-24, 13:47
Hi.
Yes everything is wired network. Only thing between pi an Xp box is one switch and two patch cables, all together in one room.
For now the mount is still ok after almost 9 hours, will see tomorrow.
For the network issues. I will arrange another switch and some cables to do a test with only pi, XP box, W7 box and router on the switch connected. That said, this wil mean i have something to do with family since wifi and a lot of other things will stop working in the house.

d6jg
2017-07-24, 14:06
It is possible that the LAN card on your Pi is faulty. I had to send one back as it had no HDMI output.

castalla
2017-07-24, 14:14
It is possible that the LAN card on your Pi is faulty. I had to send one back as it had no HDMI output.

Other mounts (eg. La Cie) seemed stable ... problems only with win shares. Why not suspect the switch?

d6jg
2017-07-24, 14:21
Other mounts (eg. La Cie) seemed stable ... problems only with win shares. Why not suspect the switch?

True.
I'd go to minimal setup and add stuff back in until it breaks.
It definitely isn't to do with pCP and the way it mounts that's for sure.

castalla
2017-07-24, 14:29
True.
I'd go to minimal setup and add stuff back in until it breaks.
It definitely isn't to do with pCP and the way it mounts that's for sure.

Agreed.

atrocity
2017-07-24, 14:41
Absolutely. What would be the result, I assume it is supposed to do a reconnect?

I'd expect to see one of two things: Either it just silently remounts the drive or it gives a (hopefully meaningful) error message. The Ubuntu bug I referenced would give "host down" message or something along those lines.

Assuming it actually remounts, maybe you could use cron to issue the command every few minutes or something. Not very elegant, but if it works...

atrocity
2017-07-24, 14:46
Other mounts (eg. La Cie) seemed stable ... problems only with win shares. Why not suspect the switch?

I've had two different switches over the years that managed to fail in ways that would screw up one piece of equipment but not another. In both cases, I blamed the affected equipment until further failure made me replace the switches, whereupon the "broken" clients magically worked again.

Mnyb
2017-07-24, 20:56
Did you try to browse the file system with a terminal window not just LMS ?

LMS expect that the assigned music folder is always there when it's running it's does not like when music directories disappear.

edwin2006
2017-07-25, 00:41
Did you try to browse the file system with a terminal window not just LMS ?

LMS expect that the assigned music folder is always there when it's running it's does not like when music directories disappear.
How do i browse the file system with terminal window?

For the network. That's why i'm getting switch etc for this weekend to test. Weird thing is that PcP is now running for almost 20 hours and mount to XP share is still ok. Also the radio function show's all item's (i.e. staff recommended, local, music, sports etc)

For LMS and missing files. That seems the course why LMS is getting so up said when browsing file structure. It entirely put's the pi down. After long wait i had a few times i had to pull the power plug.

For the remount with cron. If it gets that far i would appreciate some help in setting this up.

For the lan card of the pi. How can we test this?

bpa
2017-07-25, 02:29
If it's a network related issues - have a look at network interface stats at both ends - for wired connections there should be no (at leasts very very few) errors, timeouts or retransmissions. TCP Resets only occur "normally" when a link goes up/down.

When drive drops off Pcp are there no messages in the log file or dmesg ?

edwin2006
2017-07-25, 02:36
When drive drops off Pcp are there no messages in the log file or dmesg ?

Where to find these in the pi (still learning linux)?
Did do a brief wireshark capture and did not noticed. Will look better this weekend.
btw. Now almost 22 hours later and XP share still ok.

bpa
2017-07-25, 03:43
Where to find these in the pi (still learning linux)?
Did do a brief wireshark capture and did not noticed. Will look better this weekend.
btw. Now almost 22 hours later and XP share still ok.
I don't use Pcp so some command may not be available/same name

Dmesg shows "kernel " stuff so it usually show more "interesting" low level events. Try the dmesg command at a shell prompt. To identify the message you are looking for - I'd force a similar event to get activity logged in demsg. For example, disconnect network cable from Windows system - this will should down drive in PCP in an uncontrolled manner as disabling a Windows drive or Wiundows network interface might end up in a controlled dismount.

System log: syslog is the log of "normal" system events and so less interesting when looking for odd behaviour, is usually kept in a directory like /var/log

For interface TCP statistics try "netstat -s" - you need to do snapshots and then compare values of interesting TCP counters such as TCPRetrans or IP "failed" and "dropped" counters

edwin2006
2017-07-25, 09:31
28 hours alter en XP share still ok. Wil be out of town for the next few days. I'll keep the pi running and one squeezeplayer plying an internet station and report back when i'm home again.
Thanks so far for all the help and support!

paul-
2017-07-25, 09:40
ifconfig will show you data on the interface for packets in/out retries.

syslog is not enabled by default. Add syslog to cmdline.txt to enable (or use the Bootcodes web page In the Extras area )

One thing with pcp, there is no daemon watching the mounts. So if cifs cannot manage the connection at the kernel level, you will have to write your own service. As others have said, if music folders disappear LMS is not going to be happy. So you would have to shut down LMS, remount your disk, restart LMS I have cifs connections to a NAS that are running for weeks upon end, with the NAS going to sleep and waking up, and pCP handles that fine.

cifs expects a response when it dismounts. So if you network is dropping or your remote server is dropping, cifs may not get a response. The default timeout is quite high (I want to say 5 minutes) so that may lead you to believe that it is "locking" up the pi.

One last thing, you may want to install cifs-utils.tcz extension, it contains a few more cifs specific tools, they are normally not needed, but may help.

edwin2006
2017-07-27, 08:15
pi with lMS running for more than 3 days, mounting point to XP share still ok.
Realize now that when to issue started I reduced the number of connected SB's from 6 to 2 (one pi and one SB Radio). Since the connection now seems to be stable. Perhaps that's also a direction to investigate by adding the rest of the SB's. i.e connecting them back to the network. Especially because the family veto'd me out for network changes this weekend.

edwin2006
2017-07-29, 02:20
More than 4 day's working now. I've now added 1 boom playing internet radio. So in total 1x pcp, 1x radio and 1x boom connected to LMS pi. No network changes due to family veto this weekend.

drmatt
2017-07-29, 06:44
More than 4 day's working now. I've now added 1 boom playing internet radio. So in total 1x pcp, 1x radio and 1x boom connected to LMS pi. No network changes due to family veto this weekend.Sheesh do you have to submit change requests to the family to work on the network...?! :)


Transcoded from Matt's brain by Tapatalk

slartibartfast
2017-07-29, 07:27
Sheesh do you have to submit change requests to the family to work on the network...?! :)


Transcoded from Matt's brain by TapatalkI guess they want it to work.[emoji3]

Sent from my SM-G900F using Tapatalk

edwin2006
2017-07-29, 07:57
I guess they want it to work.[emoji3]
Sent from my SM-G900F using Tapatalk
Right, especially this weekend :(

edwin2006
2017-07-30, 07:23
so far share is holding ok. Will add a touch into the game and see what happens.

edwin2006
2017-07-30, 23:50
Still working ok. Will add an SB2. So now continu playing:
1x boom, 1x radio, 1x touch, 1x Pcp, 1x SB2
Boom, Radio and PcP via wifi, Touch & SB2 wired. LMS and XP share also wired.

edwin2006
2017-08-01, 01:22
Still running ok for 7 days. Realised that at the moment of loosing the xp share connection i had also every player streaming a different stream.
So now i'm running 1x boom, 1x radio, 1x touch, 1x Pcp, 1x SB2 with al different internet stream's. It's is weird that everything still running ok (no offence to pcp or lms btw).
I can not stand it that the the problem can't be reproduced because now i can't figure out what was causing it in the first place. Will run current config for a day or so and than add the last player (PcP3.20)

edwin2006
2017-08-02, 07:10
9 days so far, share still connected.
Will add the last PcP with another internet stream. So LMS is now controlling 6 player's with all different internet stream's.
Let's see.

d6jg
2017-08-02, 13:09
If it's working why worry any further. It could have been something as simple as a dodgy Ethernet cable which you have inadvertently fixed by swapping things around. It did sound like a networking issue.
The first rule of networking is "check the cables".
The second rule is also "check the cables"
You'd be surprised how many times a loose cable has been the cause of a problem in my experience.

edwin2006
2017-08-03, 09:17
Perhaps your right.
For now everything is rock solid and running for 10+ days.
Everyone thanks for thinking on this :)

edwin2006
2017-08-15, 07:50
Just an update.
Up until a power outage last wednesday (august 9th) everything kept on running so really not a clue what went wrong before.
After the power outage i got a corrupt SD card and had to reinstall PcP and LMS. (and 2 other pi's also an corrupt SD card)
Now prefs and cache on usb stick on the LMS pi and all other non-pcp pi's on UPS.