|
They are not fooling around with security in FreeBSD. I just finished installing the 8.0 RC and am currently left without root access to the box I had to take several steps to even enable SSH and things like that. Fine. But I did not expect that I would not be able to 'su' to root without adding myself to a particular group. And since I can't get to root, I don't have the permissions to do the fix I found. Of course I can drag a monitor and keyboard over to the thing (again) and login there as root (I hope!) and fix it but WOW these guys are careful. If you get your FreeBSD box remotely rooted you have to go out of your way to let someone have even the oppuortunity! roadhead fucked around with this message at 02:27 on Oct 8, 2009 |
# ¿ Oct 8, 2009 02:18 |
|
|
# ¿ May 14, 2024 17:57 |
|
Bob Morales posted:Dare I ask what the 20 hard drives are for? Only hooking up 10 right away - but to answer your question I want a very large Raid-Z2 pool for everything I have currently on optical media. The other 10 bays are for when I want to put in a second Raid-Z2 and mirror the first one! adorai posted:he's probably building a home virtualization environment and needs 20 spindles worth of IO. The machine has a Phenom II x 3 705e with 4 gigs of DDR2-800 on a MA785G-UD3H. I still need a few more PCI-E 4 and 2 port SATA cads to hook all 20 bays up actually, right now only the first 8 are hooked into the anything besides a fan-out cable. I won't actually be storing anything interesting, except maybe all the home-grown ripped from my Sony Digital8 camera that will never ever see the light of day! adorai posted:requiring users to be in the wheel group in order to su to root from ssh has been in freebsd for a long time. I always think it's odd when I install redhat or centos and don't have to be in the wheel group, in fact, I can just ssh in as root! Forgot to mention this is my first Foray into BSDLand roadhead fucked around with this message at 16:21 on Oct 10, 2009 |
# ¿ Oct 8, 2009 03:07 |
|
Got myself some wheels! Installing mencoder - FreeBSD has a high "scrolling text" in response to commands ratio, all I typed was "make" and the bitch has been going crazy for a bit now.
|
# ¿ Oct 8, 2009 22:56 |
|
Anyone code:
|
# ¿ Oct 30, 2009 14:19 |
|
This is a fun one. I was having trouble with an in-place upgrade, so decided to do a fresh 8.0RC2 install since my previous RC1 install was my first BSD install ever, and I now had much better idea of the packages I would actually be requiring. In my due diligence I export my main Raid-Z2 "zpool export storage" - thinking this would make it easier to correctly import the pool into the new install. I could not have been more wrong. The installs are wigging out early on, with the screen getting all weird and never prompting me. I blame this on the dodgy IDE dvd-rom I've been using, and just go back to the still untouched RC1 install (I was trying the clean install on a different drive) - and I try and import my array. Invalid vdev configuration! HURRAY! I try -f. Same. All the discs show online, but the pool just can NOT be re-imported. One thing that might factor in here, is when the pool was created, I had /dev/ad20 and /dev/ad22 (in addition to 6-18 by even numbers) - at some point I apparently did something in the BIOS to cause 20 and 22 to become 3 and 4. Strangely the ZFS didn't flinch, I didn't even have to re-silver, it just kept on trucking. It was in this state of vdev labels or whatever when I exported it. I have since the data loss investigated the option I changed, and can now have the drives appear as either set of names with a reboot and a trip into the BIOS. That doesn't help. I've pretty much resigned myself to total data loss at this point (it WAS the back-up, so I have to go about the task of filling it again. A lot of what I backed up to it is still on DVD-R) but have not yet created a new pool with the disks because I haven't spent enough time troubleshooting it (none really) and there is the possibility someone might still be able to help me So uhhh guys. Whats your prognosis? Have I wasted 4 weeks of time that I spent uploading data onto this server, or can this thing be re-imported?
|
# ¿ Nov 1, 2009 22:05 |
|
SamDabbers posted:Have you tried booting off of an OpenSolaris live CD? It might be able to mount your zpool where the FreeBSD port of ZFS wasn't able to. It's not a solution by any stretch, but it would tell you if your zpool is corrupt or if you've encountered a particularly nasty bug. Good idea, heading to grab the ISO now
|
# ¿ Nov 2, 2009 13:25 |
|
If anyone remembers my problem (exported Raid-Z2 pool under 8.0RC1 - never to be able to import again) I finally got opensolaris booting (got a new IDE optical drive in) and it has an even lower opinion of the pool that FreeBSD does.Opensolaris posted:jack@opensolaris:~# zpool import 8.0RC1 posted:hydra# zpool import I think the devices are just being enumerated incorrectly - so the labels written to the disks during export don't match what its finding now. Is there anyway to edit/move these around manually? Solaris just thinks 7 of the devices are in the wrong place (I think?) or maybe the drives really are all rear end end up? roadhead fucked around with this message at 00:29 on Nov 10, 2009 |
# ¿ Nov 10, 2009 00:20 |
|
SamDabbers posted:The import action takes into account that the disks may not have the same /dev nodes as they did when they were exported. The ZFS labels written to disk don't contain the /dev paths of the disks when the zpool was exported. They contain, among other metadata, the name/UUID of the zpool and the UUIDs of the other member disks. As long as all member disks are present in the system when you go to import the zpool, ZFS should be able to figure out the stripe order of the zpool from the labels. This is what I get for using ZFS ported to a release candidate I guess I booted into 8.0RC1 just after looking at it in Solaris to get that report, why are they of such vastly differing opinion as to the health of the individual drives in the array?
|
# ¿ Nov 10, 2009 01:32 |
|
SamDabbers posted:It looks like they both have the same opinion of the state of the zpool, but the OpenSolaris implementation (which is a few revisions ahead of the FreeBSD port) gives you more detail as to which drives contain corrupt data. I've resigned myself to building a new array and transferring those 3 TB back over the network again, but I really wish I had left the pool alone rather than exporting it cest la vie.
|
# ¿ Nov 10, 2009 14:51 |
|
Just got one of these to replace a 16 GB CF card and one of these - after my first CF to IDE adapter BLEW OUT A TRACE on the PCB. The Transcend SSD is here, but I haven't used it yet as after putting my CF card in the new adapter, FreeBSD 8.0 kept on trucking, and the array was unphased - pretty amazing considering the hard-drive with the OPERATING SYSTEM on it essentially disappeared out from under it, causing a dirty shutdown once I noticed its LED's were off, though both my SSH sessions had already become unresponsive and it would no longer respond to a ping. Can I add the Transcend SSD with the 16GB CF drive in some sort of strange Raid-1 - half the 32 gig SSD in the RAID and half un-used or formatted for swap?
|
# ¿ Dec 8, 2009 15:19 |
|
w_hat posted:I can't believe an export caused that. I'm definitely sticking with OpenSolaris now. After extensive testing it looks like you are right. While trying to track down stability issues that ultimately I determined to be caused by using 4 2 gig DDR2 DIMMs, I also accidentally unlocked the 4th core on my Phenom II 705e, and I believe it was during these shenanigans that I hosed up the array. Also while that 4th core was active nothing on the box worked quite right So the box is great now, albeit with 4 gigs of RAM instead of 8, and FreeBSD/ZFS is awesome.
|
# ¿ Jan 4, 2010 17:20 |
|
code:
code:
|
# ¿ Mar 24, 2010 14:48 |
|
jandrese posted:mountd is for servicing new nfs mount requests, it can be safely restarted, and in fact probably should since it appears to have gotten itself stuck. Thanks, restarted it and it appears to be consuming a normal amount of CPU time now.
|
# ¿ Mar 24, 2010 16:24 |
|
SmirkingJack posted:I was mostly looking for software suggestions. I didn't look at the handbook too closely since it looked like a general DNS guide and more of a Bind how-to, but I went back and set it up. As it turns out, it was far less scary and complicated than I thought it would be. Thanks, everyone, for the suggestions! Thats because *NIX and Bind ARE the general use DNS
|
# ¿ May 28, 2010 16:40 |
|
I have a RaidZ2 Zpool with 10 devices, Western Digital Green Drives. Today I decided, in my infinite idiocy to bring the box down and run WDTLER on the drives, as currently the array would stall long enough to cause a kernel panic at times, and I thought this would be a fix. 9 drives have no problems, but WD-WMAVU0467050 says "can't be set" - I think WD removed the ability to change this from newer drives, as the remaining 9 all have much lower serial numbers. With all but 1 drive changed to a TLER of 7 seconds (read and write) I boot back into FreeBSD 8. And I can't mount, import, or really do anything with my array except see this: code:
So can changing the tler setting of a drive change it enough for Zpool to wig out?
|
# ¿ Jul 7, 2010 00:28 |
|
netmazk posted:I can't recall to what degree FreeBSD's ZFS implementation relies on disk signatures instead of drive numbers (we use glabels on our ZFS arrays), but its possible that you shuffled the drives around and it is "confused" as to what is what. I didn't move the drives around, in fact I didn't move the drives at all. Everything is in the same place on the same ports, I just ran wdtler on them, thats the only difference. I can't export code:
|
# ¿ Jul 7, 2010 01:02 |
|
FISHMANPET posted:Well you can't export it because it's not imported. Although I can't see why it thinks your vdev is incomplete. That being said, if you don't give a poo poo about any data on the pool, just create a new vdev. And if I care "great shits" about all the data on the pool? Argh, its all backed up, but none of it is on-line or capable of being pulled faster than 100 megabit (at best), so I would love to save this one if possible, since its over 5 TB of stuff Also "zpool status" says something about not being able to initialize ZFS lib when I run as a non-root user, I remember being able to run zpool before as a non-privileged user, is it because its not mounting or un-related? ZDB on FreeBSD says code:
roadhead fucked around with this message at 01:40 on Jul 7, 2010 |
# ¿ Jul 7, 2010 01:32 |
|
Bob Morales posted:Wouldn't that be a bad idea? I changed the head-parking timing a long time ago, and I believe the array even rebooted and was no worse for the wear. How changing the TLER is different I am presently un-aware. Everything was hunky-dory until I tried to make it "better"
|
# ¿ Jul 7, 2010 01:42 |
|
netmazk posted:Try running zdb -l /dev/ad4. Repeat for each device. Each drive should successfully display 4 labels which all contain essentially the same information describing the zpool. Should I be piping these into files for later perusal, or is the act of querying for the info what I'm after? ad20 was unique with its Label 2 and Label 3 "failing to unpack" I didn't notice that for any of the other devices, is that bad ? code:
|
# ¿ Jul 7, 2010 03:03 |
|
FISHMANPET posted:I only brought it up because you said you tried to destroy the pool, so I assumed there wasn't any data on there. Last time I had troubles with a pool, "destroying" then "zpool import -D" was what fixed it, I don't think its ever really "destroyed" until the disks are assigned to new pools and silvered or something right? Also I just noticed that the serial number of ad20 is the drive that would not accept new TLER settings (467,xxx and the rest of the drives are 1xx,xxx) - since its a Raidz2 I should be able to lose this drive COMPLETELY (and another!) and still re-silver onto one of my spares right? roadhead fucked around with this message at 04:32 on Jul 7, 2010 |
# ¿ Jul 7, 2010 04:02 |
|
netmazk posted:Sounds pretty logical, and the zdb output makes it seem like it's the culprit. Have you tried to 'zpool detach' the drive? I don't think it'll let you if the pool isn't imported, but who knows. If it doesn't, I would shutdown, remove the suspected drive, and start it back up and see if it helps. It can't really hurt at this point. FYI, the second two labels (the ones that are missing) are stored at the end of the disk, rather than the beginning. code:
code:
Import finished, and I got this code:
roadhead fucked around with this message at 05:23 on Jul 7, 2010 |
# ¿ Jul 7, 2010 05:08 |
|
netmazk posted:I would take a minute and do a quick wipe of your old "ad20" in another machine. Then plug it back in and treat it just like you are replacing a dead drive with a brand new one. Yea I tried putting the most recent incarnation of ad20 in there, and it went back to its old "degraded" status :/ Rebooting again without anything in that bay. So any drive I put in there needs to have the "ZFS Smell" cleaned off first? With a drive totally missing, I think it even imported on boot this time, I was able to do a "zfs mount storage/stuff" and even get an LS of that filesystem, so everything should be fine as soon as I can get this thing re-silvered onto one of these spares, which I will do tomorrow. roadhead fucked around with this message at 05:33 on Jul 7, 2010 |
# ¿ Jul 7, 2010 05:29 |
|
roadhead posted:Yea I tried putting the most recent incarnation of ad20 in there, and it went back to its old "degraded" status :/ Ok When I have a freshly NTFS formatted drive connected up as /dev/ad20 I get UNAVAIL - code:
if I just SLIDE IT OUT with the machine powered on and everything, I at least get DEGRADED status - code:
However I did load the drive up in my gaming rig prior, and after running tler-on, I booted into Win 7 did a GPT/Quick NTFS format. Did I need a full format? Is there something else I can do to start the re-silver manually?
|
# ¿ Jul 7, 2010 23:44 |
|
Anyone want to tell me what happens if I export the pool in this degraded state, reboot; plug ad20 back in so it can be detected, and then try to import the pool? Should I be exporting on every shutdown?
|
# ¿ Jul 8, 2010 00:04 |
|
netmazk posted:I'm not sure what that will do. I don't think its going to help. I would use DD to write /dev/zero to the first 100MB of the disk just in case the format didn't reach far enough to catch the second label (I have no idea how big they are...). ok trying to rid this drive of all its labels, it had a complete set as far as I could tell. The first few dd attempts only got the early ones. code:
Ok I pulled AD20 and well... code:
roadhead fucked around with this message at 03:30 on Jul 8, 2010 |
# ¿ Jul 8, 2010 03:20 |
|
netmazk posted:What happens if you boot up with the drive in place now? Does the zpool come up or go back to being unavail? If it comes up, just do the replace. With a drive plugged into that port I still get UNAVAIL. However the drive still can't shake these two labels. code:
|
# ¿ Jul 8, 2010 04:03 |
|
feld posted:8.1 has been out since 7/20. I've already upgraded a few servers. Don't see any issues so far anyway Anyone got tips for upgrading to 8.1 from 8.0, with a custom kernel? code:
|
# ¿ Jul 23, 2010 17:17 |
|
Just updated to version 0.5.3 of SABnzbdplus, (http://www.freshports.org/news/sabnzbdplus/) and fixed a couple of package problems with cherrypy by reinstalling, but it still can't launch and gets the following error : code:
EDIT: Ok if you go and clean out /usr/local/lib/python2.6/site-packages/sabnzbd/* Manually, THEN make reinstall clean the port, it works fine. Will remember this in the future roadhead fucked around with this message at 21:45 on Aug 28, 2010 |
# ¿ Aug 27, 2010 17:00 |
|
code:
Freaks me out not noticing this sooner. No more letting SSHD listen on the default port!
|
# ¿ Oct 11, 2010 16:53 |
|
Bob Morales posted:There are other ways to go about it other than changing the default port. You'll just get scanned and then they'll try that port anyway. You would be right with a dedicated attack, but I think these are lazy Chinese hackers with a collection of pilfered user/pass combos just hammering every box out there that answers on port 22. Expanding that to scanning all ports on every IP for listening services multiples the required traffic about 65,536 times. I'll definitely be checking the auth.log more often, and if they do bother to discover the new port, additional measures will be taken.
|
# ¿ Oct 11, 2010 17:21 |
|
Which Java run-time is the best? I use PS3 Media Server on my box and presently its on diablo-jdk1.6.0 - but I thought I heard somewhere that OpenJDK was better? Anyone have an opinion?
|
# ¿ Oct 12, 2010 16:04 |
|
Finally getting around to trying to actually use OpenVPN - and I can load pages served by the BSD box across the link, other machines on the LAN can ping the VPN IP of the server. But I can't ping other machines on the LAN with the client, or ping the client from the LAN. I added a route to my gateway to direct 192.168.254.0/8 traffic to 192.168.1.2 - the local IP of the BSD box. I can ping 192.168.254.1 from either side, LAN or VPN - but not say 192.168.254.6 which is the IP my client is getting. I can ping 192.168.1.2 from the VPN, but not 192.168.1.1 or anything else. Must be a routing/firewall thing I've yet to configure eh? UPDATE: Crashed Apache on the gateway using the web interface to change the metric on the only static route I've put on the device. Telnet in and view the routes, and suddenly it decides to work. I guess it wasn't fully set/needed a bump. Of course this was a problem with the device in the equation running Linux! EDIT: Ok I can ping, and FileZilla will SFTP (over VPN, seems like overkill!) but won't vanilla FTP. I can use windows CLI FTP. I have also set the DNS IPs, tested them using nslookup, and can get DNS resolution via nslookup, but not just using ping at a cmd prompt. Could this be a problem with my config? Its an IP Tunnel using UDP. EDIT2: Discovered how to push DHCP options like the DNS suffix to my windows clients, and DNS resolution is working great now. I much prefer http://camera/ to http://whatever.dyndns.org:8080/ - but I am just weird I guess roadhead fucked around with this message at 15:38 on Jan 17, 2011 |
# ¿ Jan 10, 2011 19:20 |
|
This is why you setup several "datasets" under the one volume, each can have different ZFS options, such as which hash to use, compression and what level, and lots of other stuff.
|
# ¿ Jan 17, 2011 17:38 |
|
Marinmo posted:1,5TB Seagate Barracuda 7200rpm, so it's not the green edition stuff that's hampering me. All the datasets pull from the same pool of free disk space, its just that you can have different options for each one depending on the kind of data you are storing. Look at my free space for instance code:
|
# ¿ Jan 18, 2011 22:45 |
|
conntrack posted:I had two axe devices that simply burnt out after being online for a month. After it happened two times i just gave up the usb plan. I thought GigE performance on my re0 device was bad (no jumbo frame support in the FreeBSD drive) - but at least its stable!
|
# ¿ Feb 18, 2011 15:53 |
|
complex posted:8.2-RELEASE came out on the FTPs on Sunday. Please use a mirror. 'freebsd-update fetch' would choose a mirror at random, correct ?
|
# ¿ Feb 22, 2011 23:28 |
|
IanMalcolm posted:Guys, I have a ZFS problem here. I'm currently running FreeNAS 0.7, and the new 8-RC1 version came out recently with a new ZFS version. In the readme they say that it is not possible to upgrade an existing zpool, doing so would destroy the data. That is truly strange, both my pools went through zpool upgrade with all their data intact. However this is FreeBSD 8.1->8.2 and I believe ZFS version 14->15 - Not sure why FreeNAS would delete all your data...
|
# ¿ Mar 8, 2011 17:00 |
|
Ok I finally got around to the rest of the steps on the way to finishing up my 9.0 upgrade. I did the "shutdown -r now" part (really the thing that made me keep putting it off, didn't want to lose the up-time!) and started the "portupgrade -af" part earlier today... on a single user session... when I am ALWAYS OTHERWISE USING TMUX. The reason I wasn't using tmux is because well - uhhh i was re-building all my ports and was going to do another "freebsd-update install" soon so I thought a single actual SSH session would be fine. Of course this compiling went on ALL DAY TODAY. And when 5 rolled around I stupidly just shut my notebook (with the putty session on it) like I always do, cause hey, tmux right? When I got home and ssh in from my desktop it hit me - I am an amazingly stupid human being who should not be allowed to play with the switches and knobs. I killed my shell in the middle of a "portupgrade -af" surely at least leaving some half-compiled port somewhere on my system. So uhhh - will it start all over at the beginning you think or be able to pick up where it left off?
|
# ¿ Mar 21, 2012 00:06 |
|
Xenomorph posted:It just appeared on update3/4/5.freebsd.org! I probably should have known this and been ready, but upgrading to 9.1 from 9.0 wiped out all the "state" in transmission-daemon. All the BSD ISOs I was helping seed have to be added back manually and re-verified. BSD is super easy to admin but I always forget the little things since I don't generally have to do anything TO the box besides use it 99.9% of the time.
|
# ¿ Dec 12, 2012 22:59 |
|
|
# ¿ May 14, 2024 17:57 |
|
Hey over the weekend 2 of the GPT devices in my ZFS Raid-Z2 decided to disappear. Sort of. I can still pull the smart info from these drives with their adaxx designation, but their (label?) in /dev/gpt/ is missing. And 'zpool status' shows the following - code:
On the other hand with the two problem children failed out that particular array is faster than its been in a long time! Two fresh drives is probably the right answer here eh? 'gpart show' output appears to be missing for the two devices in question. I apparently did not write down the correct /dev <-> gpt translations. Great. 'dmesg' says - code:
roadhead fucked around with this message at 23:44 on Jan 7, 2013 |
# ¿ Jan 7, 2013 22:14 |