Manual update fails / ramdrive - bug?
doing a manual update via uploading the *.tgz (payling around tih 2.2 rc) caused an error.
indicating write error / disk full.
pfsense is installed as a vm (inside vmware vsphere)
2 vcpu (1s/2c)
ramdrives for both /tmp and /var
(doubled recommended values: /tmp 80mb, /var 120mb)
when i increased the /tmp ramdrive-size to 400mb (before 80mb) the error was gone and the update finished successfully.
conclusion: the tgz-file is uploaded to /tmp (which didn't have sufficent space for the 80mb+ tgz not to think about the unpacked file)
either a check when update is invoked or a hint with the configuration option to use ramdrive would be great!
besides: thx a lot to pfsense team - very good job !!!
we have some 20+ pfsenses running with our customers (between 10 and 150 users) for quite a time - started with pfsense 2.0
both physical and virtual - very few problems - quite stable!
some general thoughts/observations:
all virtual pfsense-installations have no ! swap-partition - never had any problems with memory
(mostly 1gb ram, no squid or stuff like this (= no excessive file-/disk-usage))
an embeddded image (for cf-card/usb-stick) running in a virtual machine would be great (thus embedded installs are designed to minimize disk-access - a virtual firewall with fs-corruption is quite a problem if the systems doesn't come up again after vmware failover - or another kind of hard crash)