56 minutes ago, ken-ji said:

+1 As I've noticed mine to be slowly growing as well.

 

Please provide more details:

- which version are you running?

- what app is running?

- how are you monitoring memory usage

- anything else you think might be relevant

 

We are in a data-collection phase since there is nothing obviously wrong in the code.  This is necessary to help steer where to look for the issue.  thx

@limetech Sorry, about that.

I've following this and the original thread.

 

* Version 6.4.1

* Apps: transmission, Embyserver, nginx, ELK, mariadb

* Monitoring using top, by running this for a while in a terminal. (I have 16G of RAM)

Quote

# top | grep shfs

23975 root      20   0 8399532 7.520g    836 S   0.7 48.4 169:42.28 shfs       
23975 root      20   0 8399532 7.520g    836 S   0.3 48.4 169:42.29 shfs       
23975 root      20   0 8399532 7.520g    836 S   0.3 48.4 169:42.30 shfs       
23975 root      20   0 8402408 7.522g    836 S  19.5 48.4 169:42.89 shfs       
23975 root      20   0 8405672 7.525g    836 S  26.2 48.4 169:43.68 shfs       
23975 root      20   0 8409832 7.529g    836 S  29.2 48.5 169:44.56 shfs       
23975 root      20   0 8413284 7.533g    836 S  30.2 48.5 169:45.47 shfs       
23975 root      20   0 8414016 7.534g    836 S   6.0 48.5 169:45.65 shfs       
23975 root      20   0 8414016 7.534g    836 S   1.7 48.5 169:45.70 shfs       
23975 root      20   0 8414016 7.534g    836 S   0.3 48.5 169:45.71 shfs

...
23975 root      20   0 8414016 7.534g    836 S   0.3 48.5 169:45.94 shfs       
23975 root      20   0 8414016 7.534g    836 S   0.3 48.5 169:45.95 shfs       
23975 root      20   0 8414016 7.534g    836 S   0.3 48.5 169:45.96 shfs       
23975 root      20   0 8414016 7.534g    836 S   0.7 48.5 169:45.98 shfs       
23975 root      20   0 8414016 7.535g    836 S   0.3 48.5 169:45.99 shfs

Note the sharp rise of the memory, which happened while I did a

Quote

tree -p /mnt/user/appdata

which is consistent to the current findings by LT

 

* as a data point - all my apps are using /mnt/user/appdata rather than /mnt/cache/appdata

 and I have this sysctl settings set setup ever since: 

Quote

# cat /boot/config/sysctl.conf 
vm.dirty_ratio=2
vm.dirty_background_ratio=1

Also, running the drop_caches didn't do anything.

4 hours ago, ken-ji said:

I've following this and the original thread.

 

Please try 6.5.0-rc1.  We added some more logging capability which will be available in -rc2.

I sent limetech diagnostics with shfsExtra=-logging 2.

 

Now, I'm testing 6.5.0 rc1 and it's working fine, no leaks at the moment, in fact, after sending the message to @limetech .... it went from 6400 to 5776 ... looks like is working as intended

I'm heading out for a couple of weeks, so won't be able to do any diagnostics, but I have the same problem since moving to 6.4.0. Logs attached to my post HERE ..

Edited March 4, 20188 yr by Simon

32 minutes ago, Simon said:

I'm heading out for a couple of weeks, so won't be able to do any diagnostics, but I have the same problem since moving to 6.4.0. Logs attached to my post HERE ..

 

Known issue.  Please try latest 6.5.0-rc

Only an idiot would patch to a RC build 10 minutes before leaving to the airport for a two week vacation. So that's what I just did.

Edited March 4, 20188 yr by Simon

23 minutes ago, Simon said:

Only an idiot would patch to a RC build 10 minutes before leaving to the airport for a two week vacation. So that's what I just did.

 

But you're supposed to setup remote desktop connection and bring a laptop so you can still work on vacation - that's what I have to do

1 hour ago, limetech said:

 

But you're supposed to setup remote desktop connection and bring a laptop so you can still work on vacation - that's what I have to do

Yeah, imagine being the family's IT guy. But wait... that's what a lot of us here are.

On 3/4/2018 at 2:24 PM, limetech said:

 

But you're supposed to setup remote desktop connection and bring a laptop so you can still work on vacation - that's what I have to do

OK, I VPN'd in and it's still up almost 8 days later. Thanks.

Yeah, imagine being the family's IT guy. But wait... that's what a lot of us here are.

I actually compiled a DVB build using my phone, via VPN and SSH whilst on honeymoon...

Sent from my LG-H815 using Tapatalk

@limetech I'm unable to upgrade to 6.5.0rc to see if the shfs problem has gone away, due to me being away, but between the observed very slowly growing memory usage in my case and the relative lack of apps running, I can restart the array and free up the memory for at least a few weeks, probably months.

None of the people reporting this problem earlier in the thread have weighed in with results from a recent 6.5.0-rc 

 

I see only one report of success with 6.5.0-rc1 from cferrero

 

Djoss, Jeronyson, vanes, megagator, ffiarpg, Leliil, matthope?

 

i am on RC6 now! memory usage is not growing anymore

39 minutes ago, dabl said:

None of the people reporting this problem earlier in the thread have weighed in with results from a recent 6.5.0-rc 

 

No news is typically good news.

13 hours ago, dabl said:

None of the people reporting this problem earlier in the thread have weighed in with results from a recent 6.5.0-rc 

 

I see only one report of success with 6.5.0-rc1 from cferrero

 

Djoss, Jeronyson, vanes, megagator, ffiarpg, Leliil, matthope?

 

I didn't have the chance to try 6.5.0-rc yet....

On 13/03/2018 at 6:13 PM, dabl said:

None of the people reporting this problem earlier in the thread have weighed in with results from a recent 6.5.0-rc 

 

I see only one report of success with 6.5.0-rc1 from cferrero

 

Djoss, Jeronyson, vanes, megagator, ffiarpg, Leliil, matthope?

 

 

I'd rather wait for the 6.5 stable before upgrading the OS. The memory leak is quite slow on my side, about 2Go per week. It's more an annoyance than a serious issue for me.

 

EDIT1: I'm trying the 6.5 stable at the moment, I will keep you updated.

 

EDIT2: Not an issue anymore. 6.5 seems to fix it.

Edited March 22, 20188 yr by matthope

2 minutes ago, matthope said:

 

I'd rather wait for the 6.5 stable before upgrading the OS. The memory leak is quite slow on my side, about 2Go per week. It's more an annoyance than a serious issue for me.

 

6.5 Stable is released.

On 3/13/2018 at 7:15 AM, ken-ji said:

@limetech I'm unable to upgrade to 6.5.0rc to see if the shfs problem has gone away, due to me being away, but between the observed very slowly growing memory usage in my case and the relative lack of apps running, I can restart the array and free up the memory for at least a few weeks, probably months.

So far so good , seeing as memory usage of shfs is actually decreasing at times... but I guess time will tell... its kinda too early to conclusively say the leaks have been plugged.

EDIT: sorry, forgot to state that I did upgrade to 6.5.0 and it seems ok so far.

Edited March 17, 20188 yr by ken-ji

Is it confirmed that there was an error?

Hi!

Im not sure where to put this in but I’ve got some problems with my server. When I go to the WebUI of UNRAID I see at the bottom that the array is undefined. On any page is this (Screenshot).

 

Does anyone know how to fix it?

 

Thank you!

I suspect this means that the flash drive has dropped offline for some reason.   We would need the system diagnostics to confirm this.

yes i saw this, this was when stick was bad.