Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
lilbean
Oct 2, 2003

ior posted:


Holy poo poo that's a nice case. I'm going to try to track one of them down in Toronto and replace my current (also Debian) system with it.

Adbot
ADBOT LOVES YOU

lilbean
Oct 2, 2003

If you type "zfs get all" you'll get a list of supported properties for a filesystem - check if CIFS is there.

lilbean
Oct 2, 2003

amerrykan posted:

Thanks, lilbean. *high five, shmup buddy*

I ran 'zfs get all > getall.txt', then opened it with gedit. Nothing about CIFS, but I see 'mypool/storage' has sharesmb set to "on".

From Windows, 'ping server-vm' responds with the correct IP, but 'net view server-vm' fails like no service is listening (System error 53). Other shares work, of course.
Hm, take a look at the sharemgr tool and its man page maybe. I've done NFS a lot on Solaris 10, but never CIFS so I'm not sure of the specifics but it could be that the filesystem was marked to be shared but not actually shared because of an error.

lilbean
Oct 2, 2003

Stonefish posted:

If I understand JBOD correctly, it just strings the disks together, end to end.

[Disk1 ][Disk2 ][Disk3 ]
[FILE-A][FILE-B][FILE-C][FILE-D][FILE-E][FILE-F][FILE-G][FILE-H]

If you lose disk 2, you lose files C, D, E and F, and that's once you recover the data from 1 and 3.

I think.
A lot of controllers will use the term JBOD to mean each disk is presented to the OS separately with no on-controller RAID too.

lilbean
Oct 2, 2003

Stonefish posted:

A lot of controllers are loving stupid.

(I had spaces in the "ascii art" up there, but they went away. Fuckit. /code)
Well, personally I'd rather have something named a JBOD that let me do RAID in the host systems software than doing concatenation in hardware :)

lilbean
Oct 2, 2003

Stonefish posted:

You mean you want a "SATA Interface"? Yeah, I want one of those too.

Seriously. I want a PCIE 8x card with 8 SATA ports on it. No loving raid, or at least the cheap kind that can be completely disabled and doesn't generate a fuckload of heat.

Nobody loving makes these. I'd be cool with someone taking the generic SATA chip from a mobo and slapping it on a PCIE card, as long as it had at least 8 SATA ports on a single card. Use two chips if you must. Full speed, no bloody splitters.

I like me some linux kernel raid, but it's bloody hard to implement with a significant quantity (~16) of disks.
It's not a want, the SANs I buy at work are all JBODs that just give me access to the disks, and I use MD/LVM or ZFS on them (3310s, J4200s, etc). They all offer hardware RAID but ZFS is so much slicker. Sadly I can't afford any of that poo poo at home though.

Looking around though, are cards like this what you want?
http://www.cooldrives.com/8-channel-8-port-sata-pci-card.html

lilbean
Oct 2, 2003

adorai posted:

It should soon be available in raid-z (zpools/zfs/zwhatever new things sun is doing)
It's been available in any of the ZFS-backed storage pools for a while, as well as in the previous Sun implementation of volume management (SVM/DiskSuite).

lilbean
Oct 2, 2003

Sock on a Fish posted:

Huh, I just figured it'd be like expanding a filesystem in a VM after you expand the virtual disk. Thanks for the tip.

Also doesn't ZFS only allow expandability of a pool, but not the individual components in the pool? i.e., you could add a raid-z set to a pool, but you can't add a disk to the raid-z set.
Yep, that's correct. You can add new virtual devices to any pool, and replace any physical devices in pools, but you can't add brand new devices to a RAIDZ or RAIDZ2 virtual device.

lilbean
Oct 2, 2003

WickedMetalHead posted:

I think the official recommendation for ZFS is multiple raidz(2)'s in a pool instead of one big raidz(2) anyways. I might be wrong on that but i am pretty sure that is what sun recommends, and how they have the thumper configured.
Yep, that's bang on. On mine, it's split into RAIDZ2 vdevs comprised of 6 disks each. And it's loving awesome.

lilbean
Oct 2, 2003

gregday posted:

Just destroyed and rebuilt my ZFS pool. I was stupidly using four 1TB disks with no mirrors, which I believe would be the equivalent of running a RAID-0 "scary raid."

Now I have a 2TB pool with a mirror of each disk. But I'm trying to make sure I have a clean upgrade path. Let's say I slowly want to replace those 1TB with 2TB disks, I can just replace the mirrors one at a time, right?

There wouldn't be any issue having a mirror comprising a 1TB and a 2TB right? It's just that it will simply not be able use the full 2TB until I replace the other 1TB, correct?
You can definitely replace them one at a time, but you'll only see space increases when a pair of mirrored disks is upgraded. So if you have 4 1TB disks in two mirror vdevs and replace one disk nothing will change. But when you pop the second in, you'll have your extra terabyte (or whatever).

lilbean
Oct 2, 2003

JerseyMonkey posted:

So, I have a DNS-323 (fun_plug enabled) with (2) Seagate LP 1TB drives in a RAID1 configuration. Apparently I purchased the most unreliable drives, and now my RAID has needed rebuild after degrading. Is there anything I can use on the RAID like S.M.A.R.T. to read off reports, as I have seen one of the drives fail (yet began working after being offline for a day or so), and I suspect the other of possibly failing in the near future as well.

I am wondering how to deal with the data currently on the NAS, I am half-expecting I should purchase another 1TB drive to rebuild the RAID with, so I can send off both drives for replacement, except that seems a bit overkill. Bleh, I built this NAS to offload my data to be in a safer environment, but I somehow pick horridly unreliable drives making it a very unsafe environment.
Never used one, but it looks like there's add-ons for funplug to do SMART monitoring reports:
http://wiki.dns323.info/howto:smartd_email

lilbean
Oct 2, 2003

You created two pools. You want to create one pool comprised of the two vdevs:
code:
zpool create tank raidz /home/adam/zfstest/dev/disk1 /home/adam/zfstest/dev/disk2 \
 raidz /home/adam/zfstest/dev/disk3 /home/adam/zfstest/dev/disk4

lilbean
Oct 2, 2003

FISHMANPET posted:

Yeah, I don't know what to do with my OpenSolaris install. I'm basically just biding my time until BTRFS implements RAID-5 and I am done with Solaris. It's great in the enterprise and awful in the home.
It's awful in the enterprise too :science:

lilbean
Oct 2, 2003

FISHMANPET posted:

Sup fellow Solaris buddy
:smith::hf::smith:

Most of my Solaris problems are political, since our Solaris guy is crazy and controlling.
We've been migrating off of it this year to Linux, and only have a few machines left. ZFS was fantastic, but it's not worth everything else that comes with the OS. And gently caress SPARCs for a web shop. No idea who thought that was a good idea.

lilbean
Oct 2, 2003

LamoTheKid posted:

Any PERC experts here? This is driving me bananas.

I have the latest driver disk, and I believe I have the latest firmware update for it 5.2.2-0072.

Do I need to step through every firmware update or can I jump from 5.0.1-0030 (the current one) right to this one?
I've only upgraded PERCs via the firmware update tool provided at http://linux.dell.com but they've all been full updates to the latest version without any intermediate steps (even on older models).

lilbean
Oct 2, 2003

mcsuede posted:

Anyone in here using uShare to create a UPnP server for serving content to an xbox360? Can't get the drat thing working, guessing it's user authentication issues or something but the uShare documentation is rather lacking. NAS is Ubuntu 10.04, mdadm raid 5, etc. Using the -x switch on uShare, just isn't showing up on the xbox as a source option. My Windows machine sees it fine, but the reason I suspect auth issues is that the Windows machine is specifically auth'd (via users/groups) for the shares on the NAS while I'm not sure how to go about that for the xbox.
I had much better luck with ps3mediaserver (http://code.google.com/p/ps3mediaserver/) which despite the name works great with the 360. It'll also transcode your files on th efly if the 360 doesn't support the format, plus do other things like add subtitles.

lilbean
Oct 2, 2003

mcsuede posted:

I tried it yesterday and it just said "can't find ps3" and didn't show up in sources on the 360. Modified the 360conf as instructed, etc. Any tips?
I haven't used it since around November, so I can't comment specifically about your issue. I do remember having to use the betas as the stable releases were lacking all the features I needed.

lilbean
Oct 2, 2003

I currently backup my home server to an external 2TB disk in a USB enclosure (JMicron). The backups are about to surpass 2TB though and I'm going to order a 3TB drive. Will I have any problems with the OS (Linux) seeing the 3TB disk through a USB enclosure?

lilbean
Oct 2, 2003

jeeves posted:

Research this based upon your specific OS, but it is definitely based on the OS of where it is linked to, and not on whether or not your enclosure is new.

Like WHS runs on Win2003 and thus will never see a drive above 2.1, regardless of whether or not it is in an external enclosure.
The OS, a vanilla Debian Squeeze system, definitely recognizes 3TB disks as it is (my secondary drive is a Hitachi 3TB). I was just wondering if the low-end USB mass storage controller chipsets might have a problem with the bigger drives.

Adbot
ADBOT LOVES YOU

lilbean
Oct 2, 2003

beyonder posted:

Does anyone else have issues with Seagate 7200.12s? Had one fail on me literally two minutes ago, third one in two months.
I bought a pair of them and had to RMA one within a couple weeks.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply