NextCloud cron issue after moving data

Hi!
My daughter decided to upload all her photos and videos to our NS7 NextCloud and filled the disk. Our NS7.9 + NextCloud was running in a VM on Proxmox.
First I increase the VM disk but that wasn’t enough for her files … :frowning:
So @Andy_Wismer suggested I move the NextCloud data from /var/lib/nethserver/nextcloud to the separate data disk attached to the NS7 VM. I created a sym link to the new data location and it restarted. All seemed to work OK.

The next day I got the following error

EXEC FAILED (/usr/sbin/sendmail): Resource temporarily unavailable

Also got the following in NextCloud admin yesterday 5 days later than the data move.

Last background job execution ran 4 days ago. Something seems wrong. Check the background job settings

I also have found this in the past day:

The PHP OPcache module is not properly configured. See the documentation :arrow_upper_right: for more information.

The OPcache interned strings buffer is nearly full. To assure that repeating strings can be effectively cached, it is recommended to apply opcache.interned_strings_buffer to your PHP configuration with a value higher than 8.

The OPcache is set for 20, so I increased that to 50.

I’ve been searching for the solution to fix the EXEC error but no luck so far.

The only thing I’ve found is that this error has something to do with not enough virtual memory to fullfil cron’s job.

I’m not sure what to do to debug from there. Any suggestions welcome.
Thanks in advance.

Hi @Nuke

How much RAM / Swap does your system have? (Both VM and Hypervisor).

My 2 cents
Andy

1 Like

Hi @Andy_Wismer

Thank you for replying

I have 32GB RAM and 0 Swap on the Proxmox hypervisor. RAM usage is about 20GB on the hypervisor so I could increase the NS7 VM. There are 12 CPU cores. The hypervisor is never using more than 30% peak CPU and average is under 5%. The hypervisor:

free -h
total used free shared buff/cache available
Mem: 31Gi 20Gi 9Gi 55Mi 1.8Gi 11Gi
Swap: 0B 0B 0B

NS7 has 8GB RAM with 8GB swap. Weird, the Proxmox GUI shows no swap but I went to the command line in NS7 and get this.

free -h
total used free shared buff/cache available
Mem: 7.6G 2.5G 802M 278M 4.4G 4.6G
Swap: 7.9G 9.5M 7.9G

Hi

Please explain to me what does the Swap (not available) of the hypervisor host Proxmox has to do with the swap of a virtual machine NS7?

These are 2 completly different hosts, each with their own (allocated) RAM and Swap and (allocated) CPU.

But: 8 GB RAM is enough for a Home NS7, mine was also running with 8 GB RAM, now has 12…

If you have 4 cores allocated to NS7, that’s enough.

Another issue to keep your eyes on is IOdelay on Proxmox.

If possible, allocate (temporary) 12 or 16 GB RAM to NS7.
And 4 or 8 cores.

Test and observe the VM.

My 2 cents
Andy

Hi @Andy_Wismer

What I meant was that the Proxmox GUI would show that the NS7 VM disk0 has a swap partition or not. The Proxmox instance shows swap on the Summary page. I was just remarking that the VM summary doesn’t show the swap. But I guess it is inside the disk0 so I guess Proxmox doesn’t see it.

I have 8 cores allocated but I’ll adjust to 12GB and 4 cores, reboot and see if I get more errors.

Thanks.

Keep 8 cores!

OK will do

Having 0 swap on Proxmox is never a good idea!

I have a clients Proxmox with 128 GB RAM, never more than 80 GB used.
The box has only 8 GB swap (Proxmox pre-installed, badly!) - and this is filled up to the brim - and only really goes down when rebooting.
In my opinion, RAM should be used, not 1000% slower swap. Normally, NO swap should be used.
This is a fast machine, but I still wonder if the lack of Swap is slowing things up.
Generally, I will install a Proxmox with 16 GB swap. PBS same.

My 2 cents
Andy

It looks like the EXEC FAILED (/usr/sbin/sendmail): Resource temporarily unavailable error has stopped for now.

But I’m getting a Preauthentication failed, SSSD [KRB5 on NS7 now. Time to do a bit more research … :slight_smile:

NS7 Load is about 2.5%, RAM 3GB and 0 Swap

Probably a timing issue…

Good morning. I can confirm that this error had disappeared with the increase in RAM.

This morning I have the following:

Unmatched Entries
INFO (RANDOM_DELAY will be scaled with factor 98% if used.)

But as an info error, I’m less concerned.

Thank you @Andy_Wismer for your help.

2 Likes