Hi

 

I've been having some trouble with my Unraid 6.1.9 server over the last few weeks. It keeps randomly stalling file transfers for several seconds.

 

- It's very intermittent, some days it will happen a lot, other days it will only happen once or twice

- This happens on multiple client machines, some Windows (SMB) some Linux (NFS).

- I've tried restarting all network devices (modem, router, switch) and server and clients.

- Nothing out of the ordinary appears in /var/log/syslog or dmesg

- All SMART reports are fine.

- I've done drive read/write tests on each drive with hdparm, good speeds each time.

- I'm running several Dockers and VMs - not sure if this is related? Tried stopping them all - problem still happens.

 

any help much appreciated! i'm quite familiar with unraid and linux - but i'm running out of ideas. if anyone has any suggestions of things to try or logs to look at, especially when the problem is occurring, that would be great

 

thanks in advance

 

some system info:

UnRAID 6.1.9

M/B: ASRock - C226 WS

CPU: Intel® Xeon® CPU E3-1230 v3 @ 3.30GHz

HVM: Enabled

IOMMU: Enabled

Cache: 256 kB, 1024 kB, 8192 kB

Memory: 16384 MB (max. installable capacity 32 GB)

Network: eth0: 1000Mb/s - Full Duplex

eth1: not connected

Kernel: Linux 4.1.18-unRAID x86_64

OpenSSL: 1.0.1s

 

Maybe try installing fix common issue and see if it sees any problem

Thanks, I didn't know about that plugin!

 

But no - nothing comes up that seems related, mainly warnings about out of date plugins.

 

Plugin Update Check not enabled Highly recommended to have dynamix check for plugin updates (including for the webUINotification Settings Ignore Error

Powerdown plugin not installed Highly recommended to install this plugin. Install via Install either through Community Applications or via the instructions HERE Ignore Error

Warnings Found Suggested Fix

Share openvpn is an implied array-only share, but files / folders exist on the cache Set Use Cache appropriately, then rerun this analysis. openvpn Settings Ignore Warning

Docker Update Check not enabled Recommended to enable update checks for docker applicationsNotification Settings Ignore Warning

This plugin (Fix Common Problems) not set to auto update Auto Update SettingsRecommended to enable auto updates for this plugin to enable further problem solving / fixes Ignore Warning

Plugin dynamix.system.stats.plg is not up to date Upgrade the plugin here: Plugins Ignore Warning

Plugin dynamix.system.temp.plg is not up to date Upgrade the plugin here: Plugins Ignore Warning

Plugin openvpn_server_x64.plg is not up to date Upgrade the plugin here: Plugins Ignore Warning

MySQL docker application has its /config folder set to /mnt/user/appdata/mysql Many (if not most docker applications) will not function correctly if their appData folder is set to a user share. Ideally, they should be set to a disk share. Either /mnt/cache/... or /mnt/diskX/... Fix it here: Docker Settings Ignore Warning

Zoneminder docker application has its /config folder set to /mnt/user/appdata/zoneminder Many (if not most docker applications) will not function correctly if their appData folder is set to a user share. Ideally, they should be set to a disk share. Either /mnt/cache/... or /mnt/diskX/... Fix it here: Docker Settings

So this seems to be affecting my VMs as well.

 

I have a Windows Server VM which is experiencing the same issue - stalled disk access or extremely slow speeds, for a period of time (several seconds).

 

I guess this rules out network isses? Sounds like a disk problem maybe?

Does anyone have any other suggestions? of any logs etc to look at?

So I'm doing a parity check - speed starts at 30MB/s but then drops to 500KB/s ...

OK. It seems like restarting the array a few times has stopped the problem occurring for now -- but I know it will come back! Annoying how intermittent it is...

 

also - i now get 20 to 30MB/s parity check speed. If I start in maintenance mode and do a parity check, I get over 100Mb/s.

 

Is this normal?

 

also - i now get 20 to 30MB/s parity check speed. If I start in maintenance mode and do a parity check, I get over 100Mb/s.

 

Only explanation for the parity check to be slower with the disks mounted is if some docker/plugin or client computer is accessing them during the parity check.

That was with all dockers and VMs stopped... and no client activity I think? I can double check that though. Would there be any other explanation?

That was with all dockers and VMs stopped... and no client activity I think? I can double check that though. Would there be any other explanation?

 

None I can think of.

Ah! Yes the cache_dirs plugin was running - I disabled it and now have 100MB/s parity check.

 

I'm still expecting this slow disk speed problem to come back though... but at least that explains one thing!

Thanks

Hello again -

 

This problem is still occurring - and it's very frustrating! Transfers to both my Windows (SMB) and Linux (NFS) clients keep stalling for no reason. Nothing in UnRaid logs, nothing obvious, drive tests are fine.

 

I find that restarting the Samba daemon helps the problem (/etc/rc.d/rc.samba restart) - but only for a while, then the stalling starts again.

 

Does anyone have any further suggestions of where to look or tests I can run??

 

thanks in advance