IRC logs for #opendune on OFTC at 2010-01-14
⏴ go to previous day
00:03:12 <tb2> why would I put a swap on a whole disk .. sounds unlikely
00:03:57 <tb2> possible it is freed memory
00:04:02 <tb2> and it is partly something
00:04:58 <tb2> the problem of raw disk access ... deleted things are there, so it is a bit fiddling :p
00:05:30 <tb2> I can't boot from it .. so clearly it is nothing
00:07:28 <tb2> I have to guess the size based on what fdisk tells me .. and even that is tricky
00:07:52 <tb2> (how big is a 'block'? :p I guess 1024 bytes ...)
00:08:45 <Xaroth> hm, c = GetTest(a); debug("id: {0}", c.name); debug("nick: {0}", c.nickname); debug("real: {0}", c.realname)
00:08:51 <Xaroth> that actually works :o
00:08:58 <Xaroth> GetTest is a C# function, returning a table
00:09:12 <Xaroth> now just to get the last table (table within a table) to return properly
00:11:14 <tb2> ah, found the root of another VPS ... that is something too ...
00:13:38 <Xaroth> I wasn't updating the HashTable properly
00:14:20 <tb2> things should be zero-byted from time to time :p
00:14:23 <tb2> avoids this confusion ...
00:15:18 <Xaroth> should have gotten a sata disk that knows the new TRIM command
00:15:32 <tb2> does it really zero-space?
00:15:39 <Xaroth> it's designed for SSDs
00:16:06 <tb2> what a better advise is: never strip data over multiple disks
00:16:08 <Xaroth> SSDs are faster when filling empty data, rather than re-writing pages
00:16:15 <tb2> you will ALWAYS regret it at some stage
00:16:42 <Xaroth> and yes, while you were snowboarding I've done some reading on SSD :P
00:17:00 <tb2> I am considering using SSD for the ESXi / Xen boot
00:17:07 <tb2> USB is not reliable enough for it
00:17:14 <tb2> but using a 80GB HDD is stupid
00:17:29 <Xaroth> and you REALLY want the expensive ones
00:17:50 <Xaroth> because their chipset is soooo much more superior
00:17:58 <tb2> yes, but this is only for booting
00:18:03 <tb2> and writing some minor stuff from time to time
00:18:18 <tb2> but USB tends to be dropped from time to time, so that really is useless
00:18:25 <Xaroth> the problem where most SSDs lose speed is the re-writing of pages
00:18:41 <tb2> but for a boot OS, that is a non-issue :)
00:18:51 <Xaroth> good chipsets avoid this by writing to empty pages instead, until there are no more empty pages
00:18:56 <tb2> I wouldn't use SSD for normal servers/VPSes yet ;)
00:19:09 <Xaroth> which, if you don't have the TRIM functionaly, means that at some certain point the disk gets slower
00:19:22 <Xaroth> because then it's filled up all its empty pages
00:19:32 <Xaroth> and has to resort to the dreaded recalculating of NAND cells
00:20:21 <tb2> YES! Recovered another root ...
00:20:28 <Xaroth> which means that it does 3 times as much as it should be doing
00:21:16 <tb2> out of the 6 grubs, 2 turned out to be real data ...
00:21:30 <tb2> I was expecting 2 more, although one is a windows VPS, I couldn't care about less
00:21:38 <tb2> the other one was kind of 'important' .. it was my router VPS :p
00:21:45 <tb2> but seemly it is not one of these :p
00:22:28 <Xaroth> Grote SinterKlaas Bonus? <<< i think this spammer is a bit late with his spam
00:22:58 <tb2> so ... now I have these recovered .. or at least I know how to recover them ... still a quest for 2 more partitions ... which are much harder to detect :s
00:24:38 <tb2> the problem is that a partition table doesn't have any header
00:24:46 <tb2> so scanning for them is useless, pointless, and most likely impossible
00:25:32 <Xaroth> I take it looking for the FAT is out of the question as well?
00:25:52 <tb2> I am almost certain it should be a XFS
00:26:10 <tb2> but at 33 MB/s, it is slow progress
00:26:24 <tb2> I think I leave that for overnight .. all XFSB hits ..
00:26:35 <Xaroth> next pc upgrade I'm getting myself eSata ports for my sharkoon
00:26:46 <Xaroth> downside to that is, if i want to use the drive in the sharkoon it needs to be on at boot time :/
00:26:59 <Xaroth> fail windows with fail autodiscover of new sata disks :/
00:27:07 <Xaroth> (fail as in, it doesn't do that)
00:27:16 <Xaroth> even mac 10.6 doesn't like it :p
00:27:23 <tb2> I opened an external USB case to insert random SATA
00:27:30 <tb2> but USB2.0 is too slow :p
00:27:45 <Xaroth> for movie storage it doesn't matter too much
00:27:58 <Xaroth> I mean, when I go sleep I dump it on, leave it to write
00:28:00 <tb2> for downloads at 100mbit/s, it does :p
00:28:01 <Xaroth> when I wake up, all done
00:30:58 <tb2> XFS is a nice storage to walk through raw
00:33:20 <tb2> okay, so I take this partial result ... recovered 2 OSes .. that is something ..
00:33:28 <tb2> 1 more to go .. and hopefully at least 1 data-storage
00:33:32 <tb2> the other one is rebuildable ..
00:33:49 <tb2> (contains downloads and stuff :p)
00:33:55 <tb2> at least the software to do the downloads is intact :)
00:34:11 <tb2> 3 days of searching .. sigh ..
10:02:37 <tb2> after days of disk recovering .. I found the beginning of most XFSs ... but it turns out LVM split up some disks ......
10:02:44 <tb2> and that table of content, is on the broken disk :(
10:12:33 <Xaroth> I got ESX4 to work... in a VM on ESX3.
10:14:30 <tb2> yes, it really stinks now ... as I can only partly access the important data ...
10:16:27 <Xaroth> now to figure out a tcpdump argument list to filter out everything NOT going to internal network
10:23:47 <tb2> and I am really out of ideas ...
10:29:48 <tb2> it seems impossible to load a drive which has a badsector at the partition table
10:31:08 <Xaroth> it decodes, it installs
10:31:12 <Xaroth> Product: vSphere 4 Enterprise Plus Licensed for 2 physical CPUs (1-12 cores per CPU)
10:31:27 <Xaroth> now to see what happens @reboot
10:35:05 <Xaroth> < tb2> it seems impossible to load a drive which has a badsector at the partition table << close to, at least
10:35:22 <Xaroth> people pay cowubes of money to have those situations fixed usually
10:35:28 <Xaroth> so doable, yes, feasable, no
10:35:44 <tb2> not what I meant .. my point is: I want to read the rest of the drive
10:35:56 <tb2> I don't care that it can't read the partition table
10:36:02 <tb2> I just want to have 'sda' access
10:36:11 <tb2> and I want to be able to read what ever I can from the drive
10:36:17 <tb2> ignoring any and all bad sectors
10:37:12 <tb2> is that really too much to ask? :(
10:37:31 <Xaroth> still no traffic, it's booting up :o
10:37:54 <Xaroth> that, technically means that you can get away by installing 1 license onto multiple servers
10:38:00 <tb2> I already need to jump-start the drive, as the BIOS fails to continue when it is connected at start-up :p
10:38:11 <tb2> then why buy that one license in the first place? :p
10:38:29 <Xaroth> no wait, it does something
10:38:40 <Xaroth> 11:38:19.921608 IP 192.168.1.15.123 > 213.249.66.35.123: NTPv4, Client, length 48
10:39:47 <Xaroth> ugh, waiting for port 443 to come up
10:40:42 <tb2> problem of the USB connector is that it resets the drive completely on read failure .. so I now jacked it into a real SATA port .. still not much more luck ...
10:40:50 <tb2> it seems, I really am fucked
10:41:16 <tb2> as figuring out where the LVM put all the pieces of the disk, is not something I see as a likely solution
10:41:21 <Xaroth> and woo, no traffic so far, i'm logged in into the vsphere client (fuck knows why that client is 100+mb to install...)
10:41:42 <Xaroth> I do wonder why the hell it's checking NTP that often :/
10:42:18 <tb2> it needs to calibrate itself .. takes a while
10:43:03 <tb2> hmm .. now the SATA drive is no longer found ... now it is really damaged :p
10:43:13 <Xaroth> did you hit it with a hammer?
10:43:23 <tb2> nope .. just stressed it too often
10:43:43 <Xaroth> also, the disk platters make nice bierviltjes :P
10:43:57 <Xaroth> bit geekish, but style
10:44:51 <Xaroth> then again, you don't -have- to tell people where you got em from :P
10:45:30 <tb2> did you know a HD could power-up without the power cable in the SATA interface?
10:45:49 <tb2> clearly ... I jacked out the power, and it spinned up
10:47:05 <tb2> nope .. the drive now seems to be REALLY lost
10:52:27 <Xaroth> at least you got -something- off of it, right?
10:52:38 <tb2> the problem is two fold: I have 2 drives
10:52:43 <tb2> which clearly were in some stripped mode
10:52:48 <tb2> the first drive is now dead, completely dead
10:52:59 <tb2> the second contains parts of the data I need, but in non-linear form, as far as I can tell
10:53:09 <Xaroth> so that's also quite fucked then
10:53:17 <tb2> so I can't recover the second without a bit of data from the first .. and the first is now really really dead
10:53:43 <tb2> I feel really fucked here yes ....
10:53:46 <Xaroth> as i said, quite fucked :/
10:54:42 <tb2> I should have tried a backup when it first failed, as then at least it did something ... but I thought: maybe if it cools down, it does more
10:54:47 <tb2> but no .. over time it starts to do less
10:55:14 <tb2> now I just become abusive towards the disk
10:58:07 <tb2> k, last solution: lets freeze the drive
10:59:47 <tb2> "LVM keeps a metadata header at the start of every physical volume, each of which is uniquely identified by a UUID. Each PV's header is a complete copy of the entire volume group's layout, including the UUIDs of all other PV, the UUIDs of all logical volumes and an allocation map of PEs to LEs. This simplifies data recovery in the event of PV loss."
10:59:50 <tb2> then why can't I see that?
10:59:58 <tb2> I start to doubt it was LVM used to span over 2 disks ...
11:00:07 <tb2> or some stupid ESX4 software
11:15:16 <tb2> mibbit sends the IRC conversation to google, and google adjusts its commercials vased on it
11:16:05 <Xaroth> and google probably pays for most of the hosting costs required for mibbit
12:13:48 <tb2> k, I opened the disk ... the heads seem misaligned :(
12:14:58 <tb2> it is a bit scary to see a disk running at 7200rpm :p
12:23:01 <Xaroth> more scary if it wasn't attached properly :P
12:23:21 <tb2> okay ... I think I can forget about this ..
12:23:28 <tb2> one of the platters shows damage because of the head
12:23:47 <Xaroth> you should be more concerned of damage on the head due to platters
12:23:56 <Xaroth> platters are more damage resistant than the heads
12:24:09 <tb2> yes ... and so if I see damage on the platter, I know something is terrible wrong :p
12:24:20 <tb2> the heads seem intact, although slightly misaligned (which was easy to fix)
12:24:58 <Xaroth> plus there's now dust in the disk :P
12:25:09 <tb2> yeah, the disk for sure now is unusable
12:25:18 <tb2> you have between 1 and 4 hours to do any recovery on it when opening
12:25:54 <Xaroth> made of hard disk platters?
12:26:15 <tb2> I throw out most of them a long time ago :p
12:26:20 <tb2> I do still have the magnets
12:26:30 <tb2> people don't believe me when I say there ar eVERY strong magnets in a HD
12:27:23 <tb2> so ... I had a /home of 64GB .. but from the start, it only contains 1 GB
12:27:31 <tb2> what are the odds of me finding the rest of it ...
12:31:51 <tb2> k, approaching the other disk from scratch again .. it does show clear signs of VMFS
12:36:04 <tb2> reminds me, do you know a good backup application? I am still looking for one .. never found a good one ...
12:41:33 <tb2> @calc 0x00037a80 - 0x0002fe80
12:45:46 <Xaroth> backup application for what?
12:46:12 <tb2> I have a lot of documents and stuff ... and I always fail to find some kind of nice application which does good compression, easy lookup, ...
12:50:43 <tb2> there is just a header saying it is a VMFS system .. but nothing to indicate where stuff starts and stops
12:51:07 <tb2> or maybe it does, but at least I can't locate it :p
12:53:25 <tb2> @calc 687950822912 - 686138883584
12:53:33 <tb2> @calc 1811939328 / 1024 / 1024
12:53:40 <tb2> @calc 1811939328 / 1024 / 1024 / 1024
12:53:52 <tb2> 1.6GB of a 64GB partition .. hmm
12:55:18 <tb2> so okay .... there are 2 things I would really like to see back: my music collection, which I most likely can forget because of the size, and the IRC logs ... lets scan the whole drive for that footprint :)
13:03:23 <tb2> lol .. found where that disk is continuing .. but guessing where exactly will be next to impossible :p
13:05:24 <tb2> wasted enough time on this ...
13:11:13 <tb2> tsss, Xen cant be installed to USB! ESXi at least can ...
13:11:36 <glx> the most instable connection
13:12:56 <tb2> so now I go out and try to find myself a new HD :)
13:14:08 <tb2> Xaroth: can ESXi run RAID-1?
13:14:25 <tb2> and can you send me how you did that with that key? :p
13:38:42 <tb2> pff, now I need to remember all the configurations ... ipv6 .. eapol .. pff ...
15:23:09 *** planetmaker has joined #openDune
15:23:09 *** ChanServ sets mode: +v planetmaker
23:27:41 *** ChanServ sets mode: +v Yexo
continue to next day ⏵