Alix 2D13 - Have you upgrade to 2.3 ? Did it go okey?
-
Like @jimp posted above, when the system has booted, use console menu option 8 to drop to a shell (or SSH in to a shell) and:
/etc/rc.conf_mount_rwThat will put the disk partitions into read-write.
Then use the GUI to change to "permanent read-write".
If you get errors when saving at that point, then maybe there is actually a problem with writing to your CF card.
-
The upgrade went ok, the box is running and the firewall seems to continue to work as expected.
As for the GUI… The update email promised a "new responsive look and feel to pfSense requiring a minimum of resizing or scrolling on a wide range of devices". Turns out, the opposite is the case. People have already commented about the GUI being so slow as to be unusable on Alix, unless you mount the CF card R/W permanently. And second the fonts and UI elements have increased in size, making the pull-down menus go off the bottom of my screen. In general, the new UI makes really bad use of screen real-estate and requires more scrolling than the old for me.
Last point, the check for available updates gets stuck on "Obtaining update status" forever...
Bottom line: the box still works, but barely.
Exactly the same problem here. GUI very slow, check for available updates gets stuck on "Obtaining update status" forever and no access to page "System -> Update"
I am on this https://www.applianceshop.eu/opnsense-a10-dual-core-rack-edition.html box except with Intel(R) Atom(TM) CPU Z510 @ 1.10GHz, 1GB ram and 4GB nanobsd.
Also CF card R/W permanently problem?
From the console or SSH shell, run:
/etc/rc.conf_mount_rw
Then back out to the menu and run option 16, and then try to login and go directly to that page.
I was not able to try the solution offered by jimp as stated above because the file /etc/rc.conf_mount_rw does not exist on my box.
Did no have all these problems on 2.2.6.
Erik
-
I was not able to try the solution offered by jimp as stated above because the file /etc/rc.conf_mount_rw does not exist on my box.
If that is true then you have much bigger problems. Every install has that file.
-
Hello all !
Running this ALIX based pfsense appliance, I upgraded to 2.3 yesterday ( nanoBDS 4g )
Everything works fine so far 8)
The only drawback I noticed is that the GUI is ( sometimes ) a bit slow to respond ( but the overall lifting was much needed and is really good )
-
For those who are seeing "Device busy" or similar and unable to force the disk read-write, try the following (preferably from the console):
umount -f /cf; fsck -t ufs -y /cf; mount -f -o rw /cf
Not something we'd normally recommend but somehow it seems that just the /cf slice is getting wedged for some people at the OS level.
Once that is done, set the permanent RW flag on Diagnostics > NanoBSD.
-
I was not able to try the solution offered by jimp as stated above because the file /etc/rc.conf_mount_rw does not exist on my box.
If that is true then you have much bigger problems. Every install has that file.
I did a search for the file on my system:
admin@nas-server:/$ sudo find / -name rc.conf_mount_rw
admin@nas-server:/$Nothing was found.
Firewall seems to work normally except for the issues stated in my earlier reply.
What could I do to correct this?
Erik
-
I was not able to try the solution offered by jimp as stated above because the file /etc/rc.conf_mount_rw does not exist on my box.
If that is true then you have much bigger problems. Every install has that file.
I did a search for the file on my system:
admin@nas-server:/$ sudo find / -name rc.conf_mount_rw
admin@nas-server:/$Nothing was found.
Firewall seems to work normally except for the issues stated in my earlier reply.
What could I do to correct this?
Erik
Hmmm, interesting. In "Diagnostic -> NanoBSD" I have set the Read/Write status to rw and I have set the set the permanent RW flag and did a save. I did not got the errors like JanW did.
When I go to the dashboard I see that the NanoBSD Boot Slice = pfsense1 / ada0s2(rw). Also the system now tells me that I am on the latest version in stead of staying forever in "Obtaining update status". I can also go to page "System -> Update" now.
Also the gui seems to be a bit faster.
Problem solved or should I still worry about the missing rc.conf_mount_rw ?
[Edit 140416 13:30 CET] I also send an e-mail to the manufacturer (https://www.applianceshop.eu/?___store=en) of my appliance to see if they ran in to this problem also. Keep you posted with the results.
Erik
-
2.3 upgrade from 2.2.6 is such a nightmare.
The proxy crash frequently and causing high disk I/O on Citrix Xen Server. Can not find any available package if using proxy under System - Advance -Misc (my connection stand behind another upstream proxy).
have to reverse back to 2.2.6 and wait for another patch comming
-
[…]
Once that is done, set the permanent RW flag on Diagnostics > NanoBSD.I see a lot of comments saying that RW should be set instead of the default ( RO ) for NanoBSD : why is that ? ???
My current boot slice is in RO but I haven't seen any problems so far
-
[…]
Once that is done, set the permanent RW flag on Diagnostics > NanoBSD.I see a lot of comments saying that RW should be set instead of the default ( RO ) for NanoBSD : why is that ? ???
My current boot slice is in RO but I haven't seen any problems so far
Many CF/SD cards, especially slower ones, are dreadfully slow switching between RW and RO. When functions in the system attempt to switch it on and off a lot, that leads to huge slowdowns on such cards. Even fast cards can get sluggish. In the worst cases, when the card is moving from RW to RO, it blocks other functions.
The RO switch is a safety belt, not a requirement. It has outlived its usefulness. Likely will be removed or defaulted to RW soon.
-
The RO switch is a safety belt, not a requirement. It has outlived its usefulness. Likely will be removed to defaulted to RW soon
Yes, given the (seeming to get worse - maybe just aging of media) slowness of CF cards with recent releases, and that actually the system underneath is good about not doing unneeded writes anyway, this seems a good thing to do. It will save loads of forum posts.
-
The RO switch is a safety belt, not a requirement. It has outlived its usefulness. Likely will be removed to defaulted to RW soon
Yes, given the (seeming to get worse - maybe just aging of media) slowness of CF cards with recent releases, and that actually the system underneath is good about not doing unneeded writes anyway, this seems a good thing to do. It will save loads of forum posts.
Part of this is that the OS used to play fast & loose with filesystem actions. They completed quick but were NO-OPs and didn't properly sync all data in a safe way. I wrote volumes about that back when we first took out the dangerous patches. I suspect FreeBSD 10.3 might be even safer in how it handles remounting, leading to another slow-down, but I have not yet run an in-depth analysis as I did before. Part of the issue is from pkg needing read/write access to its databases and it can't keep them in RAM disks.
-
just want to share my experience
upgraded from 2.2.6 to 2.3 on ALIX.2D3 4 hours ago. ran into the same issue with web interface slowness and dashboard stuck on obtaining update status as well. CPU was 100% used.
change to permanent read/write and disable automatic dashboard auto-update fixed the issue. CPU usage back down to around 4%.
compared to 2.2.6, web interface seems a bit slower on a low power system like ALIX.2D3.
other than that, OpenVPN tunnel, Zabbix Agent LTS are working as expected after the upgrade. Zabbix reported host information change and Zabbix agent version changed after the pfSense upgrade.
going to try upgrading RCC-VE 2440 box later tonight.
-
For what it's worth, I had a very smooth transition from 2.2 -> 2.3. The new UI looks beautiful, and I appreciate the huge amount of new improvements that the team's worked on. Thanks to those who got this out!
-
Hello,
I did the upgrade yesterday on my home Alix, everything stopped to work, pfSense did not boot anymore.The only way I had to restore pfSense was choosing 1 (pfSense 2.2.6) as boot option during the start-up. If I keep the serial console cable connected and I boot with 2.3, the console gets full of garbage chars right after the boot selection.
Also, all the VPNs were disabled after the boot, both IPSec and PPTP, and the dashboard was completely empty.
The upgrade was made using the Auto Upgrade feature. I will try again after having disabled pfBlockerNG (looks like has issues to update the database), in the while I leave you this feedback.
If someone has some information about the console issue after the boot will be appreciated.
Thanks,
Michele -
The only way I had to restore pfSense was choosing 1 (pfSense 2.2.6) as boot option during the start-up. If I keep the serial console cable connected and I boot with 2.3, the console gets full of garbage chars right after the boot selection.
Also, all the VPNs were disabled after the boot, both IPSec and PPTP, and the dashboard was completely empty.
What garbage on the console specifically? Sounds like you had it at a wrong baud rate.
PPTP is gone, so that is disabled. But the IPsec config doesn't get touched during upgrade. The dashboard being blank also indicates some kind of issue. Trying upgrade again after first uninstalling all packages is a good idea.
-
@cmb:
What garbage on the console specifically? Sounds like you had it at a wrong baud rate.
true, even if the Alix sends at 9600bps during the POST, and the console is configured in pfSense at 9600bps.
I am upgrading now, I will update this thread ASAP! ;)
Thanks,
Michele -
Ok, after I set the serial speed to 111500 I was able to see the console.
During the 2.3 boot it hangs at:
Generating RRD graphs…Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 fault virtual address = 0x0 fault code = supervisor read, page not present instruction pointer = 0x20:0xc12a67ea stack pointer = 0x28:0xd32de6d8 frame pointer = 0x28:0xd32de6e0 code segment = base 0x0, limit 0xfffff, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 38891 (ping) [ thread pid 38891 tid 100068 ] Stopped at bcopy+0x1a: repe movsl (%esi),%es:(%edi)then it goes in boot loop.
The only way to recover was setting back the serial at 9600bps and booting with 2.2.6.
I can give the full dump from the console if necessary.
Thanks,
Michele -
It didn't work even after a full RRD clear through the web interface then rebooting, it continue stopping at "Generating RRD graphs".
For now I can only boot on 2.2.6.
Thanks,
Michele -
Upgrade is done and I'm trying to Duplicate boot slice in Diagnostics / NanoBSD.
Clicking on Duplicate and after some time getting an empty page and the following in the log:Apr 15 09:52:20 gw.lan nginx: 2016/04/15 09:52:20 [error] 43594#0: *16 upstream timed out (60: Operation timed out) while reading upstream, client: 192.168.5.61, server: , request: "POST /diag_nanobsd.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php-fpm.socket:", host: "gw.lan", referrer: "http://gw.lan/diag_nanobsd.php"I have already tried to do it twice. How I can make sure that duplication is [not] completed without reboot?
