OpenVPN Install Errors on WDR3600 V1.5

[pbix]

I recently reloaded this router with Gargoyle v1.12.0 downloaded from the site. I then attempted to install OpenVPN. Flash on this router is tight but it said it would fit and afterwards apparently successful install claimed I had 64kB of flash remaining.

However I see the following in my logread that looks ominous including out of memory errors and segmentation faults. It appears to me that the space accounting is wrong and OPENVPN take more space that it thinks.

Tue Nov 5 09:38:11 2019 daemon.warn dnsmasq[7130]: ignoring nameserver 192.168.1.1 - local interface
Tue Nov 5 09:38:11 2019 daemon.info dnsmasq[7130]: using 13 more nameservers
Tue Nov 5 09:38:36 2019 user.notice firewall: Reloading firewall due to ifup of wan6 (eth0.2)
Tue Nov 5 09:38:37 2019 user.notice firewall: Reloading firewall due to ifup of wan (br-wan)
Tue Nov 5 09:38:43 2019 kern.warn kernel: [ 1598.718668] ipt_bandwidth: timezone shift of 300 minutes detected, adjusting
Tue Nov 5 09:38:43 2019 kern.warn kernel: [ 1598.725865] old minutes west=0, new minutes west=300
Tue Nov 5 09:40:43 2019 daemon.err uhttpd[2980]: ln: /etc/rc.d/S90openvpn: No space left on device
Tue Nov 5 09:40:43 2019 daemon.err uhttpd[2980]: ln: /etc/rc.d/K10openvpn: No space left on device
Tue Nov 5 09:40:43 2019 daemon.err uhttpd[2980]: /etc/rc.common: .: line 1: can't open '/usr/share/openvpn/openvpn.options': No such file or directory
Tue Nov 5 09:40:43 2019 kern.info kernel: [ 1718.708851] do_page_fault(): sending SIGSEGV to gpkg for invalid read access from 0000004c
Tue Nov 5 09:40:43 2019 kern.info kernel: [ 1718.717302] epc = 773f431c in libc.so[77388000+92000]
Tue Nov 5 09:40:43 2019 kern.info kernel: [ 1718.722456] ra = 773f4310 in libc.so[77388000+92000]
Tue Nov 5 09:40:43 2019 daemon.err uhttpd[2980]: Segmentation fault
Sat Mar 28 15:18:27 2020 cron.err crond[7964]: time disparity of 207637 minutes detected
root@Gargoyle:~#

[RomanHK]

Flash may have succeeded, but what about saving configurations? There may be a problem.

If you have little in Flash, try extroot to USB, it should help.

[Lantis]

Yea looks a lot like out of flash errors.
One of the things about the filesystem is you write data 1 block at a time. So even using 1byte over the threshold will use the entire 48 (or whatever size) block. I'd say for this reason the space calculation is flawed.

If someone came up with a better calculation is be very interested, but I confess I don't know enough myself.

[RomanHK]

Yea looks a lot like out of flash errors.
One of the things about the filesystem is you write data 1 block at a time. So even using 1byte over the threshold will use the entire 48 (or whatever size) block. I'd say for this reason the space calculation is flawed.

If someone came up with a better calculation is be very interested, but I confess I don't know enough myself.

Yes, our Czech manufacturer is facing a problem - it uses btrfs file system and backup, but it also leads to many rewrites.

Each Flash has a certain lifetime and a number of rewrites, and if @pbix has made many rewrites, Flash may be over. But I hope it's not the problem.

[Lantis]

It's is likely only out of space, not dead flash.

[pbix]

For the record I was not trying to preserve settings. This was a clean flash.

But I will say that the reason I was flashing is that this router just out of the blue started having reliability issues. I wish there was a way to check if the flash and RAM where all OK.

So mysteries abound.

Anyway I thought I would report my experience to whoever is maintaining the plugin system.