NAS/Storage Megathread: What is this "File Deletion" You Speak of

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > NAS/Storage Megathread: What is this "File Deletion" You Speak of

«‹›847 »

hifi: Jul 25, 2012

Mantle posted:

If I run
code:
find . -type f -exec setfacl -a2 owner@:rwatTnNcCy:fi:allow /mnt/pdrive;
I still get the error about the closing ; or +. What do the {} and \ represent in your shell command?

The "{}" is replaced by the pathname of the current file and the '\' is used to escape the ';'. See here, specifically the -exec lines. I'd imagine your issue is a lack of the '\' character.

# ? Nov 2, 2012 19:38

Adbot: ADBOT LOVES YOU

# ? May 28, 2024 04:39

Longinus00: Dec 29, 2005; Ur-Quan

hifi posted:

The "{}" is replaced by the pathname of the current file and the '\' is used to escape the ';'. See here, specifically the -exec lines. I'd imagine your issue is a lack of the '\' character.

He didn't put a space between the final argument he's passing to setfacl and the semicolon either, find has no way of knowing it wasn't supposed to be part of that argument.

Wikipedia actually has a pretty extensive page about the find command and has many examples.
http://en.wikipedia.org/wiki/Find

edit*

Mantle: In case you are wondering why you have to escape the semicolon, it's because it's already used by the shell to separate commands.

Longinus00 fucked around with this message at 19:58 on Nov 2, 2012

# ? Nov 2, 2012 19:52

Longinus00: Dec 29, 2005; Ur-Quan

tehschulman posted:

It was partially an accident, but I consider it a success because it didn't destroy the data when I rebuilt the array on one of the disks . Essentially it had come down to me running fsck and going through some of the errors and accepting the changes. It appeared as though every single block needed to be moved or fixed. I didn't go through more than a few dozen sectors before I decided this was probably a bad idea and quit out.

When I looked at the mdadm -e outputs it looked like /dev/sdc3 was the only drive out of the group that didn't seem in sync (later realizing this was probably the spare in the array). I had also found this blog about recovering a RAID5 mdadm array that seemed completely lost (he tried to rebuild the array and it seemed to work).

It's been a long battle to get to this point anyway. At one point I had the RAID stable enough and mounted that I could transfer some of my files off, but during that transfer the copy operation started to fail more frequently and things are now in worse condition (unmountable).

What you're writing and what's going on don't seem to match up. That blog post talks about recreating an array because he was ~~hoping~~pretty sure that the issue wans't bad disks but an accidental unplugging. In effect he forced md into thinking that 3 of the disks were in sync and after doing that the filesystem on the array mounted again because everything was in the right place. In your case mdadm can't even seem to figure out what your raid setup is supposed to be and you can't mount your filesystem because of the metadata is trashed (possibly because md has guessed incorrectly about how the array is setup) and this is after you pulled a disk. If you really want to test his method you would create a 2 disk raid5 using whichever two disks don't have errors on them (this may require some trial and error).

I would recommend you backup the raid superblocks using dd before doing this so you can restore them if it doesn't work. Even better (and what should have happened originally since it seems like you are willing to pay a lot of money to get data off these drives, I googled and there are quotes of several hundred for raid recovery and that number gets really high if there are any hardware issues) is stopping all the drives and copying their contents into new ones and using those to restore your data. If you mess up recreating or a rebuilding, your disk images are a simple dd away. Once you fix up the array on your new drives successfully then you're free to do whatever you want with the old drives.

# ? Nov 2, 2012 20:34

Mantle: May 15, 2004

hifi posted:

The "{}" is replaced by the pathname of the current file and the '\' is used to escape the ';'. See here, specifically the -exec lines. I'd imagine your issue is a lack of the '\' character.

I am getting closer! I have escaped the ';' like so:

code:

find . -type f -exec setfacl -a2 owner@:rwatTnNcCy:fi:allow {} \;

However, now when I run that command I am getting this error:

code:

setfacl: malformed ACL: "access permissions" field contains invalid flag "t"
setfacl: owner@:rwatTnNcCy:fi:allow: Invalid argument

I am confused because my setfacl command works as intended when run as standalone. Even when I remove the "invalid" ACL flags from the setfacl command, I still get the Invalid argument error.

# ? Nov 2, 2012 20:35

Longinus00: Dec 29, 2005; Ur-Quan

Mantle posted:

I am getting closer! I have escaped the ';' like so:
code:
find . -type f -exec setfacl -a2 owner@:rwatTnNcCy:fi:allow {} \;
However, now when I run that command I am getting this error:
code:
setfacl: malformed ACL: "access permissions" field contains invalid flag "t"
setfacl: owner@:rwatTnNcCy:fi:allow: Invalid argument
I am confused because my setfacl command works as intended when run as standalone. Even when I remove the "invalid" ACL flags from the setfacl command, I still get the Invalid argument error.

Do any of your files/directories contain spaces? You may want to try quoting the {}.

# ? Nov 2, 2012 20:36

Mantle: May 15, 2004

Longinus00 posted:

Do any of your files/directories contain spaces? You may want to try quoting the {}.

Yes, many of them contain spaces. I tried "{}" as well but I got the same error messages. I also just tried testing in a sub-branch of the tree that didn't have any names with spaces, same errors.

Mantle fucked around with this message at 21:15 on Nov 2, 2012

# ? Nov 2, 2012 21:09

hifi: Jul 25, 2012

Mantle posted:

I am getting closer! I have escaped the ';' like so:
code:
find . -type f -exec setfacl -a2 owner@:rwatTnNcCy:fi:allow {} \;
However, now when I run that command I am getting this error:
code:
setfacl: malformed ACL: "access permissions" field contains invalid flag "t"
setfacl: owner@:rwatTnNcCy:fi:allow: Invalid argument
I am confused because my setfacl command works as intended when run as standalone. Even when I remove the "invalid" ACL flags from the setfacl command, I still get the Invalid argument error.

Is this freebsd? The man page for setfacl doesn't have half of those flags, and on a test file the command doesn't work for me either, and the ACl inheritance flags can only be set on a directory (you use the -type f test for files...).

# ? Nov 2, 2012 21:30

movax: Aug 30, 2008

hifi posted:

Is this freebsd? The man page for setfacl doesn't have half of those flags, and on a test file the command doesn't work for me either, and the ACl inheritance flags can only be set on a directory (you use the -type f test for files...).

Yeah, this here...you simply might not have that flag on BSD. Table 8-2 and afterward here describe the ACL flags (short form) for ZFS under Solaris, there might be some documentation specifically for BSD as well.

# ? Nov 2, 2012 21:42

jmu: Feb 12, 2004; weoo.org

Fancy_Lad posted:

Just a heads up as it seems there are several folks in this thread that have been looking for IBM M1015 cards. There are a bunch on sale for ~$75 shipped on ebay right now:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=380492923084&ssPageName=ADME:B:SS:US:1123

From reading it appears you'd need this

http://www.ebay.com/itm/IBM-ServeRa...=item5896b33931

To enable RAID 5 on the card, does that sound right?

# ? Nov 2, 2012 22:15

movax: Aug 30, 2008

jmu posted:

From reading it appears you'd need this

http://www.ebay.com/itm/IBM-ServeRa...=item5896b33931

To enable RAID 5 on the card, does that sound right?

Yeah, if it's anything like the "i-Button" thing on the Supermicro cards.

Not a huge dealbreaker if you're just buying the card to use as a dumb HBA in IT-mode though, I'd imagine. And if you wanted a HW RAID controller, there are cheaper options like PERC 5s, right?

# ? Nov 2, 2012 22:16

jmu: Feb 12, 2004; weoo.org

Honestly I have no idea. I've been on the fence about setting up something to store DVD rips for a while, keeping them all on this 2 TB USB drive isn't cutting it. What I'd like to do is just setup something in my existing Win7 tower and run Plex on it to serve a little Roku in the other room. This system is on all the time anyway and has plenty of space for disks.

From working in IT I've got this idea that I want RAID 5 drilled into my head so I can weather a disk failure.

Basically I want to know what would be a good option for several 3 TB drives (four or five drives) and I think I'd want to go the hardware RAID route and just expose the logical disk to Windows 7 running on my tower. Bad idea?

/edit

Here's my current relevant hardware
OS booted from Samsung SSD I took out of a Dell notebook
ASUS P8P67 (REV 3.0) motherboard
Intel Core i5-2500K
XFX Core Edition PRO550W PSU
Old LIAN LI PC-A08B Tower which I love.

jmu fucked around with this message at 22:33 on Nov 2, 2012

# ? Nov 2, 2012 22:28

movax: Aug 30, 2008

jmu posted:

Honestly I have no idea. I've been on the fence about setting up something to store DVD rips for a while, keeping them all on this 2 TB USB drive isn't cutting it. What I'd like to do is just setup something in my existing Win7 tower and run Plex on it to serve a little Roku in the other room. This system is on all the time anyway and has plenty of space for disks.

From working in IT I've got this idea that I want RAID 5 drilled into my head so I can weather a disk failure.

Basically I want to know what would be a good option for several 3 TB drives (four or five drives) and I think I'd want to go the hardware RAID route and just expose the logical disk to Windows 7 running on my tower. Bad idea?

/edit

Here's my current relevant hardware
OS booted from Samsung SSD I took out of a Dell notebook
ASUS P8P67 (REV 3.0) motherboard
Intel Core i5-2500K
XFX Core Edition PRO550W PSU
Old LIAN LI PC-A08B Tower which I love.

With a P67 mobo, you could run your SSD off one of the Intel SATA 6Gb/s ports, and then put 4-5 HDDs on the other 5 Intel ports and use the PCH's integrated RAID for a net cost of $0 over the drives. My D:\ drive at the moment is 2 2TB drives in RAID 1 through PCH RAID, hasn't caused any problems.

If that doesn't work for you, then you think about getting a dedicated HW RAID controller to throw in as well.

# ? Nov 2, 2012 22:41

Mantle: May 15, 2004

hifi posted:

Is this freebsd? The man page for setfacl doesn't have half of those flags, and on a test file the command doesn't work for me either, and the ACl inheritance flags can only be set on a directory (you use the -type f test for files...).

I think you just gave me a big clue. One of the problems I've been having is google returns mixed BSD/Linux results for "setfacl" so I may have been mixing up syntaxes and options. It's amazingly annoying how different setfacl works in the two environments.

# ? Nov 2, 2012 23:31

Mantle: May 15, 2004

movax posted:

Yeah, this here...you simply might not have that flag on BSD. Table 8-2 and afterward here describe the ACL flags (short form) for ZFS under Solaris, there might be some documentation specifically for BSD as well.

I'm confused. FreeNAS runs on FreeBSD, so I should refer to BSD man pages for setfacl right? I tried using the "full_set" flag as noted on the man page (http://www.freebsd.org/cgi/man.cgi?setfacl) like this:

code:

find . -type d -exec setfacl -m owner@:full_set:fd:allow,group@:full_set:fd:allow,everyone@:full_set:fd:deny {} \;

But now I'm getting the errors

code:

setfacl: malformed ACL: "access permissions" field contains invalid flag "f"
setfacl: owner@:full_set:fd:allow,group@:full_set:fd:allow,everyone@:full_set:fd:deny: Invalid argument

So anyways, I am writing out the flags in long form like this:

code:

find . -type d -exec setfacl -m owner@:rwxpdDaARWcCo:fd:allow,group@:rwxpdDaARWcCo:fd:allow,everyone@:rwxpdDaARWcCo:fd:deny {} \;

New directories are now behaving properly:

code:

# file: testdir/
# owner: root
# group: team
            owner@:rwxpDdaARWcCo-:fd----:allow
            owner@:rwxpDdaARWcCo-:fd----:allow
            group@:rwxpDdaARWcCo-:fd----:allow
            group@:rwxpDdaARWcCo-:fd----:allow
         everyone@:rwxpDdaARWcCo-:fd----:deny
         everyone@:rwxpDdaARWcCo-:fd----:deny

But new files are not:

code:

# file: Box Sync ReadMe.pdf
# owner: james
# group: team
            owner@:----DdaARWcCo-:------:allow
            owner@:----DdaARWcCo-:------:allow
            group@:----DdaARWcCo-:------:allow
            group@:----DdaARWcCo-:------:allow
         everyone@:----DdaARWcCo-:------:deny
         everyone@:----DdaARWcCo-:------:deny
            owner@:--x-----------:------:deny
            owner@:rw-p---A-W-Co-:------:allow
            group@:-wxp----------:------:deny
            group@:r-------------:------:allow
         everyone@:-wxp---A-W-Co-:------:deny
         everyone@:r-----a-R-c--s:------:allow

Shouldn't new files be inheriting the same permissions as the parent directory since I have specified the fd flags?

# ? Nov 3, 2012 01:24

movax: Aug 30, 2008

On my phone at the moment, but maybe you can't mix the one letter short forms with long verbose form in the same command?

Also I think I set either f or d, not both, maybe check my post again.

# ? Nov 3, 2012 02:14

Mantle: May 15, 2004

movax posted:

On my phone at the moment, but maybe you can't mix the one letter short forms with long verbose form in the same command?

Also I think I set either f or d, not both, maybe check my post again.

setfacl isn't throwing any errors, so I think it's ok to write it the way I have.

I am now 99% of the way to getting this working as intended now with the following commands:

code:

find . -type f -exec setfacl -m owner@:rwpdDaARWcCo::allow,group@:rwpdDaARWcCo::allow,everyone@:rwpdDaARWcCo::deny {} \;
find . -type d -exec setfacl -m owner@:rwxpdDaARWcCo:fd:allow,group@:rwxpdDaARWcCo:fd:allow,everyone@:rwxpdDaARWcCo:fd:deny {} \;

When I am creating new files and directories in the shell and via SFTP, the permissions are being inherited correctly from the parent directory.

BUT... there are still a few conditions where the permissions aren't being set correctly:

1) when I drag and drop a new file/dir in Finder over ssh via MacFusion, the files are being created with 644 and 755 permissions
2) when I drag and drop a file/dir to the share via CIFS in Finder, the files are being created with 644 and 755 permission

When I create the files in the share via Finder or create a file directly in the share via the save function, the permissions are correct! It is only via drag and drop that the permissions aren't working.

Anyone have any idea why that might be?

TL;DR Default file ACL is being set correctly upon file creation, EXCEPT when dragging and dropping. Why?

Mantle fucked around with this message at 03:22 on Nov 3, 2012

# ? Nov 3, 2012 03:16

thebigcow: Jan 3, 2001; Bully!

Mantle posted:

setfacl isn't throwing any errors, so I think it's ok to write it the way I have.

I am now 99% of the way to getting this working as intended now with the following commands:
code:
find . -type f -exec setfacl -m owner@:rwpdDaARWcCo::allow,group@:rwpdDaARWcCo::allow,everyone@:rwpdDaARWcCo::deny {} \;
find . -type d -exec setfacl -m owner@:rwxpdDaARWcCo:fd:allow,group@:rwxpdDaARWcCo:fd:allow,everyone@:rwxpdDaARWcCo:fd:deny {} \;
When I am creating new files and directories in the shell and via SFTP, the permissions are being inherited correctly from the parent directory.

BUT... there are still a few conditions where the permissions aren't being set correctly:

1) when I drag and drop a new file/dir in Finder over ssh via MacFusion, the files are being created with 644 and 755 permissions
2) when I drag and drop a file/dir to the share via CIFS in Finder, the files are being created with 644 and 755 permission

When I create the files in the share via Finder or create a file directly in the share via the save function, the permissions are correct! It is only via drag and drop that the permissions aren't working.

Anyone have any idea why that might be?

TL;DR Default file ACL is being set correctly upon file creation, EXCEPT when dragging and dropping. Why?

Does the Samba config have this set as a file mask?

# ? Nov 3, 2012 07:18

Ninja Rope: Oct 22, 2005; Wee.

Perhaps the source system supporting POSIX permissions is re/setting the permissions to match the original after a copy?

# ? Nov 3, 2012 09:08

necrobobsledder: Mar 21, 2005; Lay down your soul to the gods rock 'n roll; Nap Ghost

File drag 'n drop is not the same thing as creating a brand new file because your permissions are transferred unless masked. With a new file, it inherits from both umask and the directory above it depending upon whether your OS can override whatever you attempt to set as a user (think it's enforceable on Solaris, not so on Linux).

Try this out to see if it's something Finder's doing or on the server side:
1. Mount CIFS or MacFuse SSH remote filesystem
2. In a terminal, go to /Volumes/<yourVolumeNameThatYouMounted>
3. Find a directory
4. Copy a file into it with cp
5. Observe permissions of the newly created file on the remote system
6. Copy same file from Finder into the remote directory

Also, maybe this seems dumb but have you tried to quote the long arguments to the -exec parameter? That is -exec setfacl -m "owner@:fUckJoO::..."

# ? Nov 3, 2012 18:04

BlankSystemDaemon: Mar 13, 2009

I finally verified the production date of my Samsung HD204UI F4EG (some of which are affected by an issue if they're from earlier than December 2010), and have setup S.M.A.R.T tests like this: A short test on the 1st, 8th, 15th, 22nd and a long test on the 28th for every month of the year, regardless of the day of the week. Is this too much?
_{_{_{I did not know disks listed production date on them.}}}

# ? Nov 3, 2012 18:18

devilmouse: Mar 26, 2004; It's just like real life.

Anyone using 3TB drives in their Norco 4020/4224? Came across this post and now I'm a bit panicky over adding a few 3TB drives to mine.

http://wsyntax.com/cs/killer-norco-case/

# ? Nov 3, 2012 22:00

sleepy gary: Jan 11, 2006

devilmouse posted:

Anyone using 3TB drives in their Norco 4020/4224? Came across this post and now I'm a bit panicky over adding a few 3TB drives to mine.

http://wsyntax.com/cs/killer-norco-case/

Came here to post this. What current draw are your 3tb drives rated for? You can easily eliminate the MOSFETs as they said in the article.

# ? Nov 3, 2012 22:04

devilmouse: Mar 26, 2004; It's just like real life.

DNova posted:

Came here to post this. What current draw are your 3tb drives rated for? You can easily eliminate the MOSFETs as they said in the article.

Haven't ordered them up yet. I was just looking at my dwindling free space and was considering whether it was time for an upgrade.

# ? Nov 3, 2012 22:09

Longinus00: Dec 29, 2005; Ur-Quan

DNova posted:

Came here to post this. What current draw are your 3tb drives rated for? You can easily eliminate the MOSFETs as they said in the article.

If you look at the pictures of the backpanes [1] you can see very clearly that the FETs that burned out spectacularly were on the 5V rail so that might be the real killer, not simply the total current draw.

1.
http://wsyntax.com/cs/killer-norco-case/images/full/norco-dead-fet-1.jpg

VVV

Yea what I meant is that if you want to buy some new drives it would be prudent to know what the current draw for the 5V rail in particular was. For instance on the product sheet linked off of the WD page for their green drives they only give a max current draw for the 12V rail (at something like 1.75A).

Longinus00 fucked around with this message at 00:26 on Nov 4, 2012

# ? Nov 4, 2012 00:14

sleepy gary: Jan 11, 2006

Longinus00 posted:

If you look at the pictures of the backpanes [1] you can see very clearly that the FETs that burned out spectacularly were on the 5V rail so that might be the real killer, not simply the total current draw.

1.
http://wsyntax.com/cs/killer-norco-case/images/full/norco-dead-fet-1.jpg

Well it cannot be the total current draw. It absolutely has to be an underrating of one or both of the rails.

# ? Nov 4, 2012 00:21

joe944: Jan 31, 2004; What does not destroy me makes me stronger.

I've been meaning to upgrade my file server for some time now, but haven't really had the funds. New job and raise in the last 6 months in addition to selling an old motorcycle, so I figured I'd spend a few bucks and completely replace my FreeNAS box. It's running an AMD x2 3800+ with 4GB of RAM with a crappy JBOD storage setup. I've got a family member in need of a working PC so this will hold them over for a bit after I swap it out.

I wanted a small case and decided to go M-ATX. Here's what I ended up with:

LIAN LI PC-A04
ASUS F2A55-M
AMD A4-5300
CORSAIR XMS 16GB

I'd like to go with a 6 drive raidz2 array and am holding off on drives for another week or so. Debating between 1TB and 2TB drives of whatever brand appears least lovely at the time and hopefully run across something resembling a good deal.

Any feedback is highly appreciated.

# ? Nov 4, 2012 03:52

DrDork: Dec 29, 2003; commanding officer of the Army of Dorkness

You can get 2TB drives for only a small premium over 1TBs these days ($80-$90 for a 1TB, $110-$130 for 2TB). If you opt for 1TBs, you can get WD Reds for $90 which is by far the best 1TB option. 2TB Reds go for $130, and overall are what I'd suggest. It's a hell of a lot easier with RAIDZ to just start off with higher capacity, rather than try to add it in later--you're talking a difference of $240 ($540 vs $780) to go from 5TB to 10TB of usable space. If you want to save a few bucks you can go for cheaper drives, like WD Greens, but they're really non-optimal for a NAS, and they're not that much cheaper, anyhow.

Your actual listed choices seem pretty solid. Just make sure you get a quality PSU to go along with it (and preferably a UPS, as well, even if it's just a small one).

DrDork fucked around with this message at 06:26 on Nov 4, 2012

# ? Nov 4, 2012 06:24

joe944: Jan 31, 2004; What does not destroy me makes me stronger.

I had a corsair 650w modular PSU laying around I planned to throw in. I already do run everything on a UPS as well, although it needs replacing very soon. Could probably use an upgrade anyhow running 3 machines.

Agreed on 2TB vs 1TB.. I was thinking about picking up 7-8 1TB instead for more spindles but 2TB drives are far more cost efficient and I need space more than speed.

E: On the other hand... 3TB

joe944 fucked around with this message at 16:59 on Nov 4, 2012

# ? Nov 4, 2012 16:56

DrDork: Dec 29, 2003; commanding officer of the Army of Dorkness

RAIDZ/ZFS isn't really a speed-demon to begin with, but with 6 drives on a "real" platform like you're planning you may very well be able to saturate your gigE connection already. I'd vote 2TB's and call it a day.

# ? Nov 4, 2012 17:00

TLG James: Jun 5, 2000; Questing ain't easy

Synology is loving amazing. In a filter happy country, was able to get openVPN going over the http connection back to my apartment to set it up. Set it up in about 25 minutes.

Probably the best purchase I've made.

# ? Nov 4, 2012 17:37

Dead Goon: Dec 13, 2002; No Obvious Flaws

I hope this is the appropriate thread to post in.

I have a N40L and am trying to make a bootable USB key to load the OS from - I'm planning on trying out unRAID. The USB key is a SanDisk Cruzer.

I'm on a Mac (10.8.2) so I'm trying to follow the guide they give in their documentation here:

http://lime-technology.com/wiki/index.php/USB_Flash_Drive_Preparation#MacOS_X

When I get to the part about installing the bootable MBR on the drive i just get the following error from terminal

code:

dd: /dev/rdisk1: Invalid argument

Help!

Dead Goon fucked around with this message at 20:50 on Nov 4, 2012

# ? Nov 4, 2012 17:54

Thermopyle: Jul 1, 2003; ...the stupid are cocksure while the intelligent are full of doubt. �Bertrand Russell

In the middle of swapping out older, smaller drives for newer, bigger drives in my home server and I was impressed and geeked-out by the level of packratting I've gotten myself into.

# ? Nov 5, 2012 00:27

Gism0: Mar 20, 2003; huuuh?

Had my first scare today.. For some reason one of my drives disappeared, so I rebooted and it thankfully came back. Any reason why that might happen? Should I be worried? (I'm using ZFS on Ubuntu 12.04)

code:

  pool: Tank
 state: DEGRADED
status: One or more devices could not be used because the label is missing or
        invalid.  Sufficient replicas exist for the pool to continue
        functioning in a degraded state.
action: Replace the device using 'zpool replace'.
   see: [url]http://zfsonlinux.org/msg/ZFS-8000-4J[/url]
 scan: none requested
config:

        NAME                                 STATE     READ WRITE CKSUM
        Tank                                 DEGRADED     0     0     0
          raidz1-0                           DEGRADED     0     0     0
            ata-ST3000DM001-1CH166_Z1F10WF6  ONLINE       0     0     0
            ata-ST3000DM001-1CH166_Z1F120XJ  ONLINE       0     0     0
            ata-ST31500341AS_9VS0BQT4        UNAVAIL      0     0     0
            ata-ST31500341AS_9VS1GRSR        ONLINE       0     0     0
            ata-ST31500341AS_9VS1HMQK        ONLINE       0     0     0

After a reboot..

code:

  pool: Tank
 state: ONLINE
status: One or more devices is currently being resilvered.  The pool will
        continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
 scan: resilver in progress since Mon Nov  5 09:56:03 2012
    53.8G scanned out of 5.88T at 284M/s, 6h34m to go
    10.7G resilvered, 0.89% done
config:
 
        NAME                                 STATE     READ WRITE CKSUM
        Tank                                 ONLINE       0     0     0
          raidz1-0                           ONLINE       0     0     0
            ata-ST3000DM001-1CH166_Z1F10WF6  ONLINE       0     0     0
            ata-ST3000DM001-1CH166_Z1F120XJ  ONLINE       0     0     0
            ata-ST31500341AS_9VS0BQT4        ONLINE       0     0     2  (resilvering)
            ata-ST31500341AS_9VS1GRSR        ONLINE       0     0     0
            ata-ST31500341AS_9VS1HMQK        ONLINE       0     0     0
 
errors: No known data errors

I'll be replacing the bottom 3 drives with new 3TB disks eventually but I'm poor just now

# ? Nov 5, 2012 01:02

VictualSquid: Feb 29, 2012; Gently enveloping the target with indiscriminate love.

Dead Goon posted:

When I get to the part about installing the bootable MBR on the drive i just get the following error from terminal
code:
dd: /dev/rdisk1: Invalid argument
Help!

did you remember to write if and 'of=/dev/rdisk' instead of only '/dev/rdisk' ?

Or the other way around if you are using some unusual version of dd.

# ? Nov 5, 2012 01:07

Devian666: Aug 20, 2008; Take some advice Chris.; Fun Shoe

Thermopyle posted:

In the middle of swapping out older, smaller drives for newer, bigger drives in my home server and I was impressed and geeked-out by the level of packratting I've gotten myself into.

This is a good start but somehow this picture made me think of the hoarder than had 10 tonnes of rubbish and 55 cats (15 of them dead) removed from his house.

At home I'm thinking about adding an external drive to my time capsule to give me more time before I need to build a home storage server.

# ? Nov 5, 2012 02:42

fletcher: Jun 27, 2003; ken park is my favorite movie; Cybernetic Crumb

In NAS4Free, I configured two folders to be shared for NFS, both with settings similar to (with a different path of course, both inside of /mnt/my_vdev though):

When I do a showmount -e it lists only the most recent one that I've added. What am I doing wrong here?

edit: /etc/exports only has a single entry in it as well, even though the web interface shows two. May be a bug? I'm going to try updating to newer rev.

fletcher fucked around with this message at 07:24 on Nov 5, 2012

# ? Nov 5, 2012 06:28

Ninja Rope: Oct 22, 2005; Wee.

Tell me about the HP ProLiant N40L. How does it perform with ZFS? How many internal SATA ports does it have? Airflow good?

# ? Nov 5, 2012 07:12

DrDork: Dec 29, 2003; commanding officer of the Army of Dorkness

Ninja Rope posted:

Tell me about the HP ProLiant N40L. How does it perform with ZFS? How many internal SATA ports does it have? Airflow good?

It's the standard suggestion for the "So, you want to make a NAS but don't want to put in that much effort..." people. It's overall a very nifty little box, and quite capable as a NAS, especially considering the price.

It has 4 "normal" HDD bays that it wants you to use, and another optical bay that, with a firmware flash, you can use as a fifth, and there's still an internal USB port for USB-key booting/OSes. That's about the limit of what you can stuff in there reasonably--some people have gotten external enclosures and a SAS card, but at that point you might as well just start off with a server case and save yourself the effort. Airflow is fine, incidentally.

Its performance is actually pretty good with ZFS, all things considered. As long as you load it up with 8GB (16GB if you're feeling so inclined) of RAM, and drop in some variety of Intel Pro NIC, it won't be the N40L hardware holding you back. Now, that many people want to use the wonky CIFS implementation that sits on top of something like FreeNAS or NAS4Free...well...the performance is hit-or-miss. Some people swear they get 100MB/s transfers all the time, others struggle at 20MB/s. Personally, my setup can't decide, and sometimes it's ~100MB/s, and sometimes it's 30MB/s, no rhyme or reason. But that's not the N40L's fault.

For $300 ($350 w/RAM and NIC) it's real hard to beat the ease of setup and performance of the N40L.

# ? Nov 5, 2012 07:46

fletcher: Jun 27, 2003; ken park is my favorite movie; Cybernetic Crumb

DrDork posted:

It has 4 "normal" HDD bays that it wants you to use, and another optical bay that, with a firmware flash, you can use as a fifth, and there's still an internal USB port for USB-key booting/OSes.

I didn't even need to mess with the firmware flash for my 5th drive that sits in the optical bay.

# ? Nov 5, 2012 07:49

Adbot: ADBOT LOVES YOU

# ? May 28, 2024 04:39

UndyingShadow: May 15, 2006; You're looking ESPECIALLY shadowy this evening, Sir

DrDork posted:

It's the standard suggestion for the "So, you want to make a NAS but don't want to put in that much effort..." people. It's overall a very nifty little box, and quite capable as a NAS, especially considering the price.

It has 4 "normal" HDD bays that it wants you to use, and another optical bay that, with a firmware flash, you can use as a fifth, and there's still an internal USB port for USB-key booting/OSes. That's about the limit of what you can stuff in there reasonably--some people have gotten external enclosures and a SAS card, but at that point you might as well just start off with a server case and save yourself the effort. Airflow is fine, incidentally.

If you get the right kind of bracket and an e-sata to SATA cable, you can run 6 drives, for that fresh raidz2 feeling.

# ? Nov 5, 2012 07:54

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > NAS/Storage Megathread: What is this "File Deletion" You Speak of

«‹›847 »