[opensuse] "no valid journal superlock found" on mkfs & mount ext4
More backup woes. I let badblks run overnight to verify a newly acquired large (128G) USB and am happy with the results. Now I come to put a file system on it. First, I try creating an ext4 file system. I'm going to play with various "T" options, but lets start with the basics. I do the mkfs.ext4 taking defaults, when its done I sync and try mounting. I get this [63849.264471] JBD2: no valid journal superblock found [63849.264478] EXT4-fs (sdb1): error loading journal Eh? So I try with an explicit journal creation - "-j" and yes that makes a not about creating the journal appear on the CLI. Again I try mounting. Same error. I try combining with various "-T" options, various "-J". Same result. For completeness, even though I'm a CLI guy and, as you can see, I'm minimalist, I try yast. That seems to create a ext4FS that it mounts. I look over the yast2 logs but they aren't easy to decipher. it seems that with yast EVERY parameter is there! far from minimalist! HOWEVER ... When I unmount and reinsert that USB I'm back to the old error message about no valid journal superblock. It seems Yast isn't that great after all. Next up, I try mkfs.reiserfs with default settings. The resulting formatted USB mounts with no problem. I'm hesitant to use that for backup despite the matter of inode exhaustion since the threading of reiserfs on the USB along with the reiserfs of /home can get knotted. Take that as it DOES, I just tried it. Next up is XFS. Well XFS doesn't seem very USB-friendly. It formats, but when i try mounting it hangs. and I can't ctrl-C to get out of the mount loop. So much for using a large USB stick for backups! Any suggestions as to what might be the problem with ext4? Or XFS for that matter? I'd really like a well balanced FS, and one problem I keep running into with many situations is inode exhaustion. Its why I've turned to reiserfs and do actually use BtrFS. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Are you sure you don't have a fake usb stick? A common problem are usb sticks that only have a fraction of the printed capacity.If you reach the end of the storage device, you get redirected to the beginning. badblks is happy with it, but filesystems are not. On 12/11/2015 02:25 PM, Anton Aylward wrote:
More backup woes.
I let badblks run overnight to verify a newly acquired large (128G) USB and am happy with the results.
Now I come to put a file system on it.
First, I try creating an ext4 file system. I'm going to play with various "T" options, but lets start with the basics.
I do the mkfs.ext4 taking defaults, when its done I sync and try mounting.
I get this
[63849.264471] JBD2: no valid journal superblock found [63849.264478] EXT4-fs (sdb1): error loading journal
Eh?
So I try with an explicit journal creation - "-j" and yes that makes a not about creating the journal appear on the CLI.
Again I try mounting. Same error.
I try combining with various "-T" options, various "-J". Same result.
For completeness, even though I'm a CLI guy and, as you can see, I'm minimalist, I try yast. That seems to create a ext4FS that it mounts. I look over the yast2 logs but they aren't easy to decipher. it seems that with yast EVERY parameter is there! far from minimalist!
HOWEVER ...
When I unmount and reinsert that USB I'm back to the old error message about no valid journal superblock. It seems Yast isn't that great after all.
Next up, I try mkfs.reiserfs with default settings. The resulting formatted USB mounts with no problem.
I'm hesitant to use that for backup despite the matter of inode exhaustion since the threading of reiserfs on the USB along with the reiserfs of /home can get knotted. Take that as it DOES, I just tried it.
Next up is XFS. Well XFS doesn't seem very USB-friendly. It formats, but when i try mounting it hangs. and I can't ctrl-C to get out of the mount loop.
So much for using a large USB stick for backups!
Any suggestions as to what might be the problem with ext4? Or XFS for that matter?
I'd really like a well balanced FS, and one problem I keep running into with many situations is inode exhaustion. Its why I've turned to reiserfs and do actually use BtrFS.
Le 11/12/2015 17:46, Florian Gleixner a écrit :
Are you sure you don't have a fake usb stick? A common problem are usb sticks that only have a fraction of the printed capacity.If you reach the end of the storage device, you get redirected to the beginning. badblks is happy with it, but filesystems are not.
if it's not a hard drive but a pen drive, test it with h2testx.exe on windows or f3read/f3write on linux jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 12/11/2015 12:04 PM, jdd wrote:
Le 11/12/2015 18:04, jdd a écrit :
if it's not a hard drive but a pen drive, test it with h2testx.exe
h2testw.exe
Like many here, "I don't do windows" -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 11/12/2015 20:20, Anton Aylward a écrit :
On 12/11/2015 12:04 PM, jdd wrote:
Le 11/12/2015 18:04, jdd a écrit :
if it's not a hard drive but a pen drive, test it with h2testx.exe
h2testw.exe
Like many here, "I don't do windows"
you didn't read my previous post with f3read/f3wrie that do the exact same thing then h2testw.exe, but for linux http://oss.digirati.com.br/f3/ but it's only usefull for flash drives and I don't thibnk your problrm is related to fakes jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 12/11/2015 12:04 PM, jdd wrote:
Le 11/12/2015 17:46, Florian Gleixner a écrit :
Are you sure you don't have a fake usb stick? A common problem are usb sticks that only have a fraction of the printed capacity.If you reach the end of the storage device, you get redirected to the beginning. badblks is happy with it, but filesystems are not.
if it's not a hard drive but a pen drive, test it with h2testx.exe on windows or f3read/f3write on linux
Unhelpful. I don't do windows and search didn't find f3read/f3write. Of course I tested with dd, quite obviously, and that seemed to verify that all 128G is writeable with no hiccups. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 11/12/2015 20:23, Anton Aylward a écrit :
Of course I tested with dd, quite obviously, and that seemed to verify that all 128G is writeable with no hiccups.
I have three (3) chinese 128Gb pen drvies that are fakes of 8 Gg... some of then broke in a matter of minutes. (all where refunded) the test is simple: write as many 1Gb files as possible, with they order number in them, then read them back. If the drive is a fake, the same files come regularly at read time. No nead to be root, works on mounted drive. http://oss.digirati.com.br/f3/ I did not test fake drives for extX they all are supposed to be USB3, but I have one that works only as usb2 (write on usb3 mode crashes openSUSE), a other one, bough from amazon works, but deadly slow even with usb3... of course, none pf then are expensive ones jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Fri, 11 Dec 2015 20:23, Anton Aylward
On 12/11/2015 12:04 PM, jdd wrote:
Le 11/12/2015 17:46, Florian Gleixner a écrit :
Are you sure you don't have a fake usb stick? A common problem are usb sticks that only have a fraction of the printed capacity.If you reach the end of the storage device, you get redirected to the beginning. badblks is happy with it, but filesystems are not.
if it's not a hard drive but a pen drive, test it with h2testx.exe on windows or f3read/f3write on linux
Unhelpful. I don't do windows and search didn't find f3read/f3write.
http://oss.digirati.com.br/f3/ (seems to have troubles, google cache has the page) https://github.com/AltraMayor/f3 (Source on githup)
Of course I tested with dd, quite obviously, and that seemed to verify that all 128G is writeable with no hiccups.
Bah, not listening again. Lets say the device is fake, and has a real size of 64GB, but fakes 128GB. The program "dd" will write the 128GB happily, it will not detect the fake. That is what such special tools are for. They write markers in a relative distance from each other, and check them. Example of such tool as pseudo code: Get device size in blocks (e.g. one per reported GB) random_num=rnd(64bit) Jump to start of device Repeat loop for n blocks: Check for marker (random_num followed by block_num) Marker exists and reports a block_num smaller than loop_num: Device is Fake Abort loop, report error Marker does not exist: Write marker (random_num followed by loop_num) Jump to next block start Loop_end dd will NOT help you here. More links: https://www.google.com/search?q=f3write+linuxw - Yamaban.
On 12/11/2015 02:59 PM, Yamaban wrote:
https://github.com/AltraMayor/f3 (Source on githup)
I got that and compiled it and have been running it for over 4 hours now. Its up to 8 files it says its created and still going strong, no error reported. I don't know what else its doing that those 8 "files" represent. I'll let it run overnight. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Thank you guys, it turns out that this supposed 128G drive is only an 8G drive. Sheesh! I can get 16G drives cheaper at Staples! SECTORS ok/corrupted/changed/overwritten Validating file 1.h2w ... 2097152/ 0/ 0/ 0 Validating file 2.h2w ... 2097152/ 0/ 0/ 0 Validating file 3.h2w ... 2097152/ 0/ 0/ 0 Validating file 4.h2w ... 2097152/ 0/ 0/ 0 Validating file 5.h2w ... 2097152/ 0/ 0/ 0 Validating file 6.h2w ... 2097152/ 0/ 0/ 0 Validating file 7.h2w ... 2097152/ 0/ 0/ 0 Validating file 8.h2w ... 483752/ 1582512/ 0/ 30888 Validating file 9.h2w ... 726842/ 142400/ 0/ 0 -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/12/2015 16:01, Anton Aylward a écrit :
Thank you guys, it turns out that this supposed 128G drive is only an 8G drive. Sheesh! I can get 16G drives cheaper at Staples!
SECTORS ok/corrupted/changed/overwritten Validating file 1.h2w ... 2097152/ 0/ 0/ 0 Validating file 2.h2w ... 2097152/ 0/ 0/ 0 Validating file 3.h2w ... 2097152/ 0/ 0/ 0 Validating file 4.h2w ... 2097152/ 0/ 0/ 0 Validating file 5.h2w ... 2097152/ 0/ 0/ 0 Validating file 6.h2w ... 2097152/ 0/ 0/ 0 Validating file 7.h2w ... 2097152/ 0/ 0/ 0 Validating file 8.h2w ... 483752/ 1582512/ 0/ 30888 Validating file 9.h2w ... 726842/ 142400/ 0/ 0
make it refunded if you can. it's a recurrent problem. Two years ago it was 32Gb, now 128Gb jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 12/12/2015 12:35 PM, jdd wrote:
make it refunded if you can.
In process.
it's a recurrent problem. Two years ago it was 32Gb, now 128Gb
Are you saying that the threshold (or what actually works) moves? Can we assume that 32G sticks and chips are the new baseline and they do work, even when advertised as 128G? Incidentally, I ran f3fix and reset it to an 8G device and put a file system on that. What is it that determines how large a USB device _appears_ to be? -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/12/2015 20:42, Anton Aylward a écrit :
Are you saying that the threshold (or what actually works) moves? Can we assume that 32G sticks and chips are the new baseline and they do work, even when advertised as 128G?
Incidentally, I ran f3fix and reset it to an 8G device and put a file system on that. What is it that determines how large a USB device _appears_ to be?
I don't know how safe it is to use only the real size, computers may be tricked to bypass and these devices are really fragiles (three on five 128Gb pen drives I buy died before the end of the test!). What I say is than nowaday there are no more 32Gb fakes (for what I know of), the problems may be about speed, and I buy routinely fast 32Gb sd cards for around $12 (amazon), so also this problem is fading. non, as the capacity increase, fakes are found on high end cheap ones. the first 3 fakes I buy where $12 each for 128Gb, the second test was $15, and where also fakes (counterfeits Kingston - but I could only see that when received on the blister). The next one I buy from china was $30 and was working only on usb2 (crash on usb3 write, works on usb3 read) Then I could buy a working usb3 128Gb pen drive for around $35 from amazon. working, but at usb2 speed :-( I guess in 6 month from now there will be no more 128Gb fakes, but may be 512Gb fakes :-( right now fast 128Gb pen drives are more expensive than ssds... jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 12/12/2015 02:54 PM, jdd wrote:
right now fast 128Gb pen drives are more expensive than ssds...
The advantage of pen drives, as many security violations (Think: Snowdon) have shown is that the7y are pocket-able. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/12/2015 22:45, Anton Aylward a écrit :
On 12/12/2015 02:54 PM, jdd wrote:
right now fast 128Gb pen drives are more expensive than ssds...
The advantage of pen drives, as many security violations (Think: Snowdon) have shown is that the7y are pocket-able.
:-) but is this an advantage :-) and makers fight to make then ten time bigger than necessary, with unusable big boxes :-( jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 12/11/2015 11:46 AM, Florian Gleixner wrote:
Are you sure you don't have a fake usb stick? A common problem are usb sticks that only have a fraction of the printed capacity.If you reach the end of the storage device, you get redirected to the beginning. badblks is happy with it, but filesystems are not.
And how would I go about determining that? The mkfs seem to write superblocks well into the device. That write doesn't fail. I've tried rate-limiting rsync to 100kb/s as it writes to the device and that ameliorates things. Yes it blows up, then I restart it and it continues on from that point. So its not that this is a 64M device, since the first break was ay 90M use on a reiserfs formatted drive, nor at the 128M point since it sailed past that to perhaps it doesn't like something in ~/.config/google-chrome ? Anyway, luckyBackup goes just so far then the rsync task thrown an IOerror about which it says very little. -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (4)
-
Anton Aylward -
Florian Gleixner -
jdd -
Yamaban