Re: finding a tarball on a fat-less fat partition--disk editor? (fwd)
to: du reply-to: [EMAIL PROTECTED], debian-user@lists.debian.org from: [EMAIL PROTECTED] Subject: Re: finding a tarball on a fat-less fat partition--disk editor? In-reply-to: Your message of "Fri, 15 Sep 2000 19:28:34 EDT." <[EMAIL PROTECTED]> References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> Mime-Version: 1.0 > A tar file comes in blocks of 512 bytes. First is the filename and > then comes other things including the word "ustar". After this header > comes the file and then comes many more header/file combinations. So > you should see "ustar" as many times as there are files in your tar > file. > If you want all the gory details, download the tar source package and > read tar.h. yikes, I can do without the gory details :) does this mean that once I find a block of a tar, I can start extracting, even if it wasn't the middle? And now that I think of it, someone mentioned that there are bad disk editors available for linux. I just realized that I can't use the same method I sed on an ext2 on a fat (unless it grew inodes while I wasn't looking :) Or should I just start using "dd if=/dev/hda7 skip=1| tar -tvf -" and incrementing the skip until I hit something (I think these two files would be the only ones ever to be created on that partition). thanks hawk --
Re: finding a tarball on a fat-less fat partition--disk editor?
chris chryed, > On Sat, Sep 16, 2000 at 02:14:56PM -0400, [EMAIL PROTECTED] wrote: > > yikes, I can do without the gory details :) does this mean that once I > > find a block of a tar, I can start extracting, even if it wasn't the > > middle? > You want to find the first block of the tar. If you will indulge a bit > of ASCII art, a tar looks like this: > +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ > | H | F | F | H | F | F | F | H | F | F | F | F | H | F | H | F | > +---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+---+ > where H is a header and F is the file's data. Each block as I said is > 512 bytes. The header contains information like how many bytes come in > the file data. If you miss the first header you miss the first file. As usual, there's only one important file (my unfiled tax return :) Even if I miss most of them, I"m ok (though I"d really like to get them all back, of course . . .) > > Or should I just start using "dd if=/dev/hda7 skip=1| tar -tvf -" and > > incrementing the skip until I hit something (I think these two files > > would be the only ones ever to be created on that partition). > This could work if you had a recently defragmented partition or a very > small tar file. In any other case, I'm going to guess that you're SOL > and going to give up. I tried to recover data from a screwed up FAT > partition once and ended up running mke2fs on it after a day. The > problem with dd on a normal filesystem is that a file can have multiple > other files in between its start and finish. I believe that these are the only two files ever written to the disk, and I am certain that there has been nothing written after it. So I'm assuming that they're contiguous . . . > > thanks > I hope it helps, me too :) thanks, I'll try again tomorrow night . . . --
Re: finding a tarball on a fat-less fat partition--disk editor? (fwd)
> > yikes, I can do without the gory details :) does this mean that once I > > find a block of a tar, I can start extracting, even if it wasn't the > > middle? > you mean "even if it wasn't the START?", right? > the answer is yes. just verified this. yes. Thanks. > > And now that I think of it, someone mentioned that there are bad disk > > editors available for linux. > i said, they suck, not that they are bad. this means, that they are not > that simple to use as diskedit for dos and lack the one or other > interesting feature - at least the last time i looked out half a year ago. > ;-) For one-use once, i'll put up with almost anything. I assumee I only need to browse a couple of k, anyway. But what are they called? grepping /var/lib/dpkg/available for disk.*edit doesn't yield anything. > > I just realized that I can't use the same > > method I sed on an ext2 on a fat (unless it grew inodes while I wasn't > > looking :) > ?? there's a how-to that explains brute-force inode searches on ext2 partitions. I recovered some important files from another machine this way a few weeks ago. thanks rick --
Re: finding a tarball on a fat-less fat partition--disk editor? (fwd)
Oswald opined, > well ... i don't remember their names. i found some of them by searching > freshmeat.net and doing a generic web search for "linux disk editor" *DOH!* Now I feel foolish :) > or > something like that. i did not bookmark them, as they all were not very > satisfactory ... :-( Hmm, since it's a disk editor, it shouldn't be a problem that I'm using a linux disk editor to mess with a FATpartition, should it? "bits is bits" :) > on a unix-system you can search for deleted nodes - if you find any, > you have a good chance of recovering them (if the disk was not filled after > the deletion), but you'll never know, what the original name of the file > was (without guessing it by inspecting the contents). > (btw: i'm not really sure, that the allocation info in an inode is completely > preserved when it is marked as deleted (i did not read the how-to), but as > i see no reason to clear it ...) There was enough. I realised what I'd done as soon as I'd done it (rm -r * ~), with the extra space before ~ . . . I recovered one file I'd forgotten about, but not what I was after. I then ran dd | strings | grep -v to recover all of the text. (fortunately, lyx stores in text format [I've used diff/patch when editing an out of date version]). > Hi! I'm a .signature virus! Copy me into your ~/.signature, please! hmm, this got through our detectors :) hawk --
Re: Begone, vile emacs!
Ray rote, > I know that when 2.2 installs in is emacs19. But of course you know that > as soon as you remove emacs RMS sends large men to your house to break >your knees. :) I like emacs but it is big. Yes, but in the true emacs anti-unix tradtion, they don't stop at doing one thing well, but instead try to do evertything. After breaking your knees, they tweak your nose, do your laundry, wash your dishes, and dislocate your elbow :) hawk --
Re: Do you recomend to upgrade to Woody?
> Has anyone used woody for mission critical stuff? I know some people run > woody-based web servers, eg, but I haven't. I used to run the unstable branches. Prior to 1.1, this was kind of necessary :) However, somewhere arround 98, iirc, it became impractical. With regular updated, I could pretty much count ona show-stopper problem every 4-6 months--the kind that would take the machine out of action for most fot the day. Mission critical? I was trying to complete a dissertaiton, so yes :) hawk --
kernel 2.4 fixes bad connection with eepro
This has come up a couple of times, including by me. I've install 2.4, and it seems to solve the flaky connection problem from the eepro card in my machine (i'd previously needed a keepup script to force-reload the network every few seconds. hawk --
Re: kernel 2.4 fixes bad connection with eepro
jessi jumbled, > actually, last time i checked, the 2.4 kernel had the same problem. i am > hoping pre9 fixes them. Typically, with pre8 (AFAIK) if you throw load at > the machine, the eerpro drivers "wig" out. This is from the debian "test-5" version. I can't stay up more than a couple of minutes with 2.2, but it's stayed up all day [then again, it used to stay up all day, at least under vary light load, and stayed up all day under FreeBSD] Unfortunately, I didn't scroll far enough down to find parallel printer support, so that module is just now compiling. Say, have I mentioned lately that the kpkg tools are a royal pain? Somewhere after an hour or so of messing with it, I always give up. And I have *yet* to succeed in getting it to avoid "make clean" even after rtfm'ing and trying to use the options the fm suggested . . . the only time it's ever been even vaguely worth the effort was when I had a fast machine available (but not on my desk) and needed a custom kernel for a 486 laptop . . . hawk --
PPP failure on the startup (serial line is looped back)
Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii I've managed once more to get the base distribution onto the kdis machine, and I think that this one was the biggest fight I"ve had in five years of debian . . . Theres' a cute bug that *can* happen, but I'm not sure what caused it, that apparently causes extracton fo drivers.tgz to go to stdout rather than the file system (at least, I assume that that's what the garble is). And error messages still go to the the raw screen, only to be overwritten immediately by the installation program before they can be read, rather than to another vt. I don't know if this one is in the bug tracking system, as I first reported it before that system existed . . . Anyway, eventually the entire system is in, and I attempted the ppp connection. The system successfully detected my modem, but fails instantly during the dselect phase--it apparently tries to grab the packages without first starting ppp. /var/log/messages yields the following: Sep 24 12:22:03 eyry kernel: PPP line discipline registered. Sep 24 12:22:03 eyry kernel: registered device ppp0 Sep 24 12:22:03 eyry pppd[238]: pppd 2.3.11 started by root, uid 0 Sep 24 12:22:06 eyry pppd[238]: Exit. Sep 24 12:22:06 eyry pppd[241]: pppd 2.3.11 started by root, uid 0 Sep 24 12:22:08 eyry pppd[241]: Serial connection established. Sep 24 12:22:08 eyry pppd[241]: Using interface ppp0 Sep 24 12:22:08 eyry pppd[241]: Connect: ppp0 <--> /dev/ttyS1 Sep 24 12:22:10 eyry pppd[241]: Serial line is looped back. <= Sep 24 12:22:10 eyry pppd[241]: Connection terminated. Sep 24 12:22:11 eyry pppd[241]: Exit. Sep 24 12:23:13 eyry pppd[269]: pppd 2.3.11 started by root, uid 0 Sep 24 12:23:14 eyry chat[276]: abort on (BUSY) Sep 24 12:23:14 eyry chat[276]: abort on (NO CARRIER) Sep 24 12:23:14 eyry chat[276]: abort on (VOICE) Sep 24 12:23:14 eyry chat[276]: abort on (NO DIALTONE) Sep 24 12:23:14 eyry chat[276]: abort on (NO DIAL TONE) Sep 24 12:23:14 eyry chat[276]: abort on (NO ANSWER) Sep 24 12:23:14 eyry chat[276]: send (ATZ^M) Sep 24 12:23:14 eyry chat[276]: expect (OK) Sep 24 12:23:59 eyry chat[276]: alarm Sep 24 12:23:59 eyry chat[276]: send (AT^M) Sep 24 12:23:59 eyry chat[276]: expect (OK) Sep 24 12:24:13 eyry pppd[302]: pppd 2.3.11 started by root, uid 0 Sep 24 12:24:13 eyry pppd[302]: Device ttyS1 is locked by pid 269 Sep 24 12:24:13 eyry pppd[302]: Exit. Sep 24 12:24:44 eyry chat[276]: alarm Sep 24 12:24:44 eyry chat[276]: Failed Sep 24 12:24:45 eyry pppd[269]: Exit. What does "serial line is looped back" mean? Is this my problem? Does this mean that the modem is on full duplex? And if so, does anyone have a list of the dipswitch settings for a 1984 era gateway modem? :)
Unidentified subject!
Subject: Re: PPP failure on the startup (serial line is looped back) In-Reply-To: Message from John Hasler <[EMAIL PROTECTED]> of "25 Sep 2000 17:31:58 CDT." <[EMAIL PROTECTED]> References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii john jabbed, > dochawk writes: > > What does "serial line is looped back" mean? > It means that pppd is seeing its own LCP packets coming back. This usually > happens when the host on the other end wants to see more text before it > starts pppd and is echoing everything it sees while it waits for the > required string. It's happening far too fast for that--It'sa only about a second from my telling it to do this to failure (with the failure message hidden behind the blue menu screens). There's no way that it's even dialing out, let alone getting through and making contacr. > Dial up your ISP with minicom and work through the login > procedure by hand to see how it works (this is not a method of connecting: > it's just research). Post the results along with the contents of > /etc/chatscripts/provider and /etc/ppp/peers/provider. Munge passwords. Can I just install the minicom deb on the raw abase system? I thought that there was a problem with this (but it's been a few years; maybe I'm just badly mistaken . . .)) hawk
Unidentified subject!
cc: David Wright <[EMAIL PROTECTED]>, hawk Subject: Re: ppp install failing (loopback?) In-Reply-To: Message from hawk of "Tue, 29 Aug 2000 13:06:07 EDT." <[EMAIL PROTECTED]> References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> <[EMAIL PROTECTED]> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii OK, here's the new data. I can dial out with minicom with no problem, and the speaker is definitely on. PPP doesn't get as far as actually dialing, unless it's sending a command string to silence the modem. When, during post-reboot installation, I get as far as being asked what method I want to use to install (immediately after configuring ppp), I choose http. THe blue menus go away, a string of "failed to get .bbb" type messages come in white on black, and then are instantly replaced with the blue menu. plog reports: Sep 27 07:14:34 eyry pppd[225]: Connect script failed Sep 27 07:14:35 eyry pppd[225]: Exit. (this is the entire output). The last 20 lines of /var/log/messages are: Sep 27 07:11:12 eyry kernel: kernel build: 2.2.17 unknown Sep 27 07:11:12 eyry kernel: options: [pci] [cardbus] [apm] Sep 27 07:11:12 eyry kernel: Intel PCIC probe: not found. Sep 27 07:11:12 eyry kernel: ds: no socket drivers loaded! Sep 27 07:13:03 eyry kernel: CSLIP: code copyright 1989 Regents of the University of California Sep 27 07:13:03 eyry kernel: PPP: version 2.3.7 (demand dialling) Sep 27 07:13:03 eyry kernel: PPP line discipline registered. Sep 27 07:13:03 eyry kernel: registered device ppp0 Sep 27 07:13:03 eyry pppd[225]: pppd 2.3.11 started by root, uid 0 Sep 27 07:13:04 eyry chat[232]: abort on (BUSY) Sep 27 07:13:04 eyry chat[232]: abort on (NO CARRIER) Sep 27 07:13:04 eyry chat[232]: abort on (VOICE) Sep 27 07:13:04 eyry chat[232]: abort on (NO DIALTONE) Sep 27 07:13:04 eyry chat[232]: abort on (NO DIAL TONE) Sep 27 07:13:04 eyry chat[232]: abort on (NO ANSWER) Sep 27 07:13:04 eyry chat[232]: send (ATZ^M) Sep 27 07:13:04 eyry chat[232]: expect (OK) Sep 27 07:13:49 eyry chat[232]: alarm Sep 27 07:13:49 eyry chat[232]: send (AT^M) Sep 27 07:13:49 eyry chat[232]: expect (OK) while the last 20 of /var/log/syslog are Sep 27 07:13:03 eyry kernel: PPP: version 2.3.7 (demand dialling) Sep 27 07:13:03 eyry kernel: PPP line discipline registered. Sep 27 07:13:03 eyry kernel: registered device ppp0 Sep 27 07:13:03 eyry pppd[225]: pppd 2.3.11 started by root, uid 0 Sep 27 07:13:04 eyry chat[232]: abort on (BUSY) Sep 27 07:13:04 eyry chat[232]: abort on (NO CARRIER) Sep 27 07:13:04 eyry chat[232]: abort on (VOICE) Sep 27 07:13:04 eyry chat[232]: abort on (NO DIALTONE) Sep 27 07:13:04 eyry chat[232]: abort on (NO DIAL TONE) Sep 27 07:13:04 eyry chat[232]: abort on (NO ANSWER) Sep 27 07:13:04 eyry chat[232]: send (ATZ^M) Sep 27 07:13:04 eyry chat[232]: expect (OK) Sep 27 07:13:25 eyry kernel: VFS: Disk change detected on device fd(2,0) Sep 27 07:13:49 eyry chat[232]: alarm Sep 27 07:13:49 eyry chat[232]: send (AT^M) Sep 27 07:13:49 eyry chat[232]: expect (OK) Sep 27 07:14:34 eyry pppd[225]: Connect script failed Sep 27 07:14:34 eyry chat[232]: alarm Sep 27 07:14:34 eyry chat[232]: Failed Sep 27 07:14:35 eyry pppd[225]: Exit. It seems that for some reason, it doesn't get an "OK" back from the modem in response to its "AT". However, it m must be getting one earlier after ATZ, or it wouldn't get this far. hawk, still baffled.
Unidentified subject!
-- cc: Subject: my .signature is gone in nmh Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii I've noticed that my .signature is no longer included, either from the command line or when using exmh. It still seems to be there: fac13pts/0:hawk>ls -l .signature -rwx--1 hawk hawk 155 Aug 16 16:22 .signature Any ideas? hawk
no valid newsgroups again (inews)
I know I"ve been through this before, but this didn't come up on my install last year (which worked fine). I can read usenet just fine, but when I try to post, I get inews: No valid newsgroups in alt.folklore.computers --which happens to be the group I"m reading. I'm not running a newserver, but getting news from news.psu.edu, if this makes a difference. --
2.4-pre5 especially susceptible to memory overload?
I have had more system hangs in the last two weeks than in the last four years put together . . . (OK, so that make s 2 :) The first incident was loading (well, attempting to load) an excessively large file into beav (a binary file editor). 1G file, 160Mb memory. It ran out of memory, and eventually stopped swapping. This was repeatable. Yesterday, I selected two columns on star office 5.2, copied to clipboard, and then my fingers slipped as I tried to do a paste-special. Same results. Both of these were done in userland. I can't go back to 2.2, as I have an eepro network card (only stays up a couple of seconds under load). hawk
automatically setting "From:" with mh
After months of on and off poking, I'm still stuck. I've tried reading assorted fm's, but get lost along the way. I need to figure out how to get mh (or nmh) to automatically include a specified From: line whenever it mails from my account (or, even better, to get it to to include this in the template I can edit. I've foudn the setting to make myself a trusted user that *can* set from:, and I've also found a system-wide template, but I need the individual user setting. As it is, I frequently forget to set the from: line. Currently, the machine is running 24/7, so it catches replies to such mail anyway, but but it is slated to be replaced with a laptop-- so replies to such mailj could be lost in the future. thanks hawk p.s. I've set the reply-to: line, as my digests frequently get garbled on inc
Re: automatically setting "From:" with mh
Robert Rambled, > I´ve simply set it in the components: > [waldner:~/Mail] egrep Waldner *comp* > components:From: Robert Waldner <[EMAIL PROTECTED]> > distcomps:Resent-From: Robert Waldner <[EMAIL PROTECTED]> > forwcomps:From: Robert Waldner <[EMAIL PROTECTED]> > rcvdistcomps:Resent-From: Robert Waldner <[EMAIL PROTECTED]> > replcomps:From: Robert Waldner <[EMAIL PROTECTED]> THanks. THis is *exactly* what I needed. Could I talk you out of the formats of the other files? Particularly, what do I need to do to get "Subject:", "References:", etc. to come out properly. I Assume it's an environmental variable or escape sequence, but I'm not making any sense out of /etc/mh/mhl.* . . . rick
Re: vi in ssh window
> Hi, > When I ssh in to my debian potato system and try to do something like vi, > or dselect, the screen messes up. For example, the vi status line occurs > at the top of the screen, and I can't see the line the cursor is currently > on. In dselect the same type things happen, stuff gets printed on lines > previous or after where it should be. It makes it really hard to use > remotely. Is this something wrong with my term settings? Any ideas? Is your terminal variable set correctly? See what your client is emulating, and set the environment variable TERM to this type. hawk
