Is nanobsd web-based upgrade supposed to be working now?

gloomrider

Cool! ;D

This is the 4G update I see. Did you happen to notice it took 6.5 minutes to flash one slice? :o

pakjebakmeel

Cool! ;D

This is the 4G update I see. Did you happen to notice it took 6.5 minutes to flash one slice? :o

Yes, 6.5 minutes of worrying that I might need to open the case again :D The CF controllers on ALIX boards don't seem to be very fast :-) So what do I do now? Do I need to copy Slice2 to Slice1 or just leave it like this? What happens if I invoke the next upgrade? Does it then upgrade Slice1 and the set that one active and reboot?

–---------------------

EDIT:

I see we cheered a bit too soon:


pfsense:~#  fdisk

******* Working on device /dev/ufs/pfsense1 *******
parameters extracted from in-core disklabel are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

Figures below won't work with BIOS for partitions not in cyl 1
parameters to be used for BIOS calculations are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

fdisk: invalid fdisk partition table found
Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 63, size 3949281 (1928 Meg), flag 80 (active)
        beg: cyl 0/ head 1/ sector 1;
        end: cyl 845/ head 15/ sector 63
The data for partition 2 is:
 <unused>The data for partition 3 is:
 <unused>The data for partition 4 is:
 <unused>pfsense:~#  ls -l /dev/ad0*
crw-r-----  1 root  operator    0,  66 Aug 26 14:28 /dev/ad0
crw-r-----  1 root  operator    0,  67 Aug 26 22:15 /dev/ad0s1
crw-r-----  1 root  operator    0,  68 Aug 26 14:28 /dev/ad0s2
crw-r-----  1 root  operator    0,  72 Aug 26 14:28 /dev/ad0s2a
crw-r-----  1 root  operator    0,  73 Aug 26 14:28 /dev/ad0s2c
crw-r-----  1 root  operator    0,  69 Aug 26 14:28 /dev/ad0s3</unused></unused></unused>

So now fdisk shows that there is no second partition and I'm missing some stuff under /dev/ad*

darnit.. Everything still works though..

UPDATE2:

After duplicating Slice2 to Slice1 the ls -l /dev/ad0* seems okay again. Fdisk is still reporting only 1 partition and an invalid parition table?? ??? ??? ??? ???


pfsense:~#  fdisk
******* Working on device /dev/ufs/pfsense1 *******
parameters extracted from in-core disklabel are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

Figures below won't work with BIOS for partitions not in cyl 1
parameters to be used for BIOS calculations are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

fdisk: invalid fdisk partition table found
Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 63, size 3949281 (1928 Meg), flag 80 (active)
        beg: cyl 0/ head 1/ sector 1;
        end: cyl 845/ head 15/ sector 63
The data for partition 2 is:
 <unused>The data for partition 3 is:
 <unused>The data for partition 4 is:
 <unused>pfsense:~#  ls -l /dev/ad*
crw-r-----  1 root  operator    0,  66 Aug 26 14:28 /dev/ad0
crw-r-----  1 root  operator    0,  67 Aug 26 22:30 /dev/ad0s1
crw-r-----  1 root  operator    0, 101 Aug 26 14:28 /dev/ad0s1a
crw-r-----  1 root  operator    0, 102 Aug 26 14:28 /dev/ad0s1c
crw-r-----  1 root  operator    0,  68 Aug 26 14:28 /dev/ad0s2
crw-r-----  1 root  operator    0,  72 Aug 26 14:28 /dev/ad0s2a
crw-r-----  1 root  operator    0,  73 Aug 26 14:28 /dev/ad0s2c
crw-r-----  1 root  operator    0,  69 Aug 26 14:28 /dev/ad0s3</unused></unused></unused>

gloomrider

Interesting. The log output shows the update seemed to work as expected. Just curious, did the update not work, or are you just concerned with what fdisk is saying?

pakjebakmeel

The update worked fine, I'm running the 26th snapshot now. No issues whatsoever.

But the experience with this Fdisk output tells me the next upgrade will fail ::)

Could someone please briefly explain the differences between slices, partitions and the /dev/ad0* files and how they relate to each other? ???

gloomrider

@pakjebakmeel:

The update worked fine, I'm running the 26th snapshot now. No issues whatsoever.

But the experience with this Fdisk output tells me the next upgrade will fail ::)

Could someone please briefly explain the differences between slices, partitions and the /dev/ad0* files and how they relate to each other? ???

There was a new 4G snapshot posted yesterday. Have you tried it?

Edit: Oops, you're running that. FYI: /dev/ufs/pfsense0 = slice 1 and /dev/ufs/pfsense1 = slice 2. You should try running fsck on the slice you're NOT running on.


/sbin/fsck_ufs -y /dev/ad0s1a

pakjebakmeel

Got this now:


pfsense:~#  /sbin/fsck_ufs -y /dev/ad0s1a
** /dev/ad0s1a (NO WRITE)
** Last Mounted on /
** Phase 1 - Check Blocks and Sizes
** Phase 2 - Check Pathnames
** Phase 3 - Check Connectivity
** Phase 4 - Check Reference Counts
** Phase 5 - Check Cyl groups
7966 files, 276870 used, 3607270 free (4254 frags, 450377 blocks, 0.1% fragmentation)

pfsense:~#  fdisk 
******* Working on device /dev/ufs/pfsense1 *******
parameters extracted from in-core disklabel are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

Figures below won't work with BIOS for partitions not in cyl 1
parameters to be used for BIOS calculations are:
cylinders=3918 heads=16 sectors/track=63 (1008 blks/cyl)

fdisk: invalid fdisk partition table found
Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 63, size 3949281 (1928 Meg), flag 80 (active)
	beg: cyl 0/ head 1/ sector 1;
	end: cyl 845/ head 15/ sector 63
The data for partition 2 is:
 <unused>The data for partition 3 is:
 <unused>The data for partition 4 is:
 <unused>pfsense:~#</unused></unused></unused>

gloomrider

I see it now. fdisk is auto-selecting the mounted root file system which has no partition table. If you want to see the partition table:

EDIT:


fdisk /dev/ad0

Be careful and don't write anything unless you know what you're doing.

I think you'll be fine for the next snap :)

EDIT:

Here's mine:


router:~#  fdisk /dev/ad0
******* Working on device /dev/ad0 *******
parameters extracted from in-core disklabel are:
cylinders=3970 heads=16 sectors/track=63 (1008 blks/cyl)

Figures below won't work with BIOS for partitions not in cyl 1
parameters to be used for BIOS calculations are:
cylinders=3970 heads=16 sectors/track=63 (1008 blks/cyl)

Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 63, size 1949409 (951 Meg), flag 80 (active)
        beg: cyl 0/ head 1/ sector 1;
        end: cyl 909/ head 15/ sector 63
The data for partition 2 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 1949535, size 1949409 (951 Meg), flag 0
        beg: cyl 910/ head 1/ sector 1;
        end: cyl 795/ head 15/ sector 63
The data for partition 3 is:
sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD)
    start 3898944, size 102816 (50 Meg), flag 0
        beg: cyl 796/ head 0/ sector 1;
        end: cyl 897/ head 15/ sector 63
The data for partition 4 is:

jzsjr

Every time I attempt to upgrade (2gig nano file) I get a message after a bit I get a failure due to the upgrade image being larger than the partition. I remember reading earlier someone have the same issue but could not find it after searching the term. I started the tail but all I got was:
pfsense:~# tail -f /cf/conf/upgrade_log.txtWARNING: Expected rawoffset 0, found
63

NanoBSD Firmware upgrade in progress…

Installing /root/firmware.tgz.
Installing /root/firmware.tgz.
SLICE 1
OLDSLICE 2
TOFLASH ad0s1
COMPLETE_PATH ad0s1a
GLABEL_SLICE pfsense0
NanoBSD Firmware upgrade in progress...

Broadcast Message from root@pfsense.local
(no tty) at 12:40 EDT...

NanoBSD Firmware upgrade in progress...

Installing /root/firmware.tgz.
Installing /root/firmware.tgz.
SLICE 1
OLDSLICE 2
TOFLASH ad0s1
COMPLETE_PATH ad0s1a
GLABEL_SLICE pfsense0

Broadcast Message from root@pfsense.local
(no tty) at 12:40 EDT...

thanks,
Jim

gloomrider

As you can see in this thread, my first successful upgrade attempt was August 9th. Prior to that, I was just reflashing every time a new snap appeared on the server. Perhaps at some point the size of the disk images changed and I never noticed it because I was reflashing the whole CF card. This is all speculation of course, but it might be useful to reflash (this is where multiple CF cards come in handy) and wait for the next snap, attempt an update and see what happens. I haven't had any issues since 8/9 and I have updated with every new update on the snapshot server.

jzsjr

10-4 Gloomrider. I have multiple CFs and reflash regularly when an update comes out and I am unable to update via the gui or the console. I just have not had any luck with this yet. I settled on the 2gig image since it looked middle of the road and that my cf card is 4gigs. It formats just under that size so I was worried that the 4 gig image/dual partitions would not exactly fit. I thought things were looking up when you and the other gentleman were successful in upgrading but no, I am still reflashing:(

gloomrider

When was the last time/snapshot you reflashed? It would seem only manual update is working. I think auto-update is still trying to use the old embedded (AKA broken) paradigm.

Nanobsd images that I download from http://snapshots.pfsense.org/FreeBSD_RELENG_7_2/pfSense_RELENG_1_2/updates are updating fine for me.

EDIT: I wonder if the fact you're using a different size card has anything to do with it? Why don't you try a 4GB image on your 4GB card?

gloomrider

@jzsjr:

10-4 Gloomrider. I have multiple CFs and reflash regularly when an update comes out and I am unable to update via the gui or the console. I just have not had any luck with this yet. I settled on the 2gig image since it looked middle of the road and that my cf card is 4gigs. It formats just under that size so I was worried that the 4 gig image/dual partitions would not exactly fit. I thought things were looking up when you and the other gentleman were successful in upgrading but no, I am still reflashing:(

Check out this thread: http://forum.pfsense.org/index.php/topic,18805.0.html

This looks an awful lot like what you and others are reporting. I have been using "dd" under both Linux and FreeBSD to flash my 2GB Sandisk cards and I've not had any of these problems.

EDIT: It appears there are subtly different sizes of CF cards. See this post: http://forum.pfsense.org/index.php/topic,18805.msg96698.html#msg96698

sigi

Hi,

running embedded since longtime. Now I wanted to switch from embedded to nanobsd, because reflashing, resizing and move rrd's from time to time sucks. First, what I learned: no more WRAP. No problem have ALIX'es too. Also read about bigger Memory nedded. No problem, have plenty of 1GB CF.

But I spent hours to test the Update mechanism. Not one success. Fails anytime. Manual, Auto, Console. I thought I have FreeBSD experience since Years. But no Idea what is going on here.

at the moment I test from pfSense-1.2.3-1g-20090824-1854-nanobsd

Option 13
Enter the URL to the .tgz update file:

tgz?

It is http://snapshots.pfsense.org/FreeBSD_RELENG_7_2/pfSense_RELENG_1_2/updates/pfSense-1.2.3-1g-20090828-0527-nanobsd-upgrade.img.gz

nope is img.gz…

Anyway, whatever I do: ad0s2x get shot...
assuming it dd's not really the right format/file?

pfSense:~# tail -f /cf/conf/upgrade_log.txt

NanoBSD upgrade starting

dd if=/dev/zero of=/dev/ad0s2 bs=1m count=1
1+0 records in
1+0 records out
1048576 bytes transferred in 0.226607 secs (4627287 bytes/sec)

/usr/bin/gzip -dc /root/firmware.tgz | /bin/dd of=/dev/ad0s2 obs=64k
949467+8 records in
7417+1 records out
486130176 bytes transferred in 116.117005 secs (4186555 bytes/sec)
After upgrade fdisk/bsdlabel

/sbin/fsck_ufs -y /dev/ad0s2a
Can't stat /dev/ad0s2a: No such file or directory
Can't stat /dev/ad0s2a: No such file or directory

/sbin/tunefs -L pfsense1 /dev/ad0s2a
tunefs: /dev/ad0s2a: could not find special device

<quote>Embedded has historically been a second class citizen</quote>
<ironic>Now it's third class...?</ironic>

No, to be honest, there seems to be a lack of testing. Even the the graphical part "firmware upgrade in Progress" is off the Center. And stays also after failing. I can not find out if I test alpha or beta?

With over 1000 reads, this topic seems to be well watched...

althornin

Just a note, I was able to successfully upgrade a pfsense 512MB cf to the latest snapshot, using manual firmware upgrade. went flawlessly, which is good, because its a remote location (non critical, just for testing).

sigi

this means:

Download from URL or local File? Or maybe WebInterface?

From pfSense-1.2.3-512mb-20090827-1405 ?and the File was pfSense-1.2.3-512mb-20090829-0427-nanobsd-upgrade.img.gz?

I think i will try this with another size than 1gb.

gloomrider

Download file appropriate to your system from here:

http://snapshots.pfsense.org/FreeBSD_RELENG_7_2/pfSense_RELENG_1_2/updates

File is now on your hard drive.

Now update the pfSense instance using the Manual Update web interface. "System:Firmware:Manual Update"

jzsjr

I have been attempting the manual updates. I am using the Sandisk ultra II 4gig cf card and chose the the 2gig image mainly because for the longest time the 4gig image was not being updated. Now it looks like the 512 and the 4gig image are the favorites to get updated. My understanding from placing the 2gig image and a 4gig cf card was that the card was formated with 2 1gig partitions and that anything else was wasted. It would seem that the fault lies somewhere in the formatting process that the nano image uses. I will try out the 4 gig image and see how that manually updates.
thanks,
Jim

jzsjr

Okay, my 4gig CF card is only showing 1.88 gig capacity and will not reformat to original size. Is this due to the nano image and windows only seeing one partition? Is there any way to restore the card to the original size or is it permanently smaller? This probably has something to do with my updating problem.
jim