Problems with FSCK

I was getting ready to install RSync on the Empeg, when I set the drive to RW I got the unchecked file system warning. When I tried to run FSCK I got this message:


empeg:/empeg/bin# ro
empeg:/empeg/bin# rw
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
empeg:/empeg/bin# ro
empeg:/empeg/bin# umount /dev/hda4
empeg:/empeg/bin# fsck -fay /
e2fsck 1.29 (24-Sep-2002)
fsck: Is a directory while trying to open /

The superblock could not be read or does not describe a correct ext2
filesystem. If the device is valid and it really contains an ext2
filesystem (and not swap or ufs or something else), then the superblock
is corrupt, and you might try running e2fsck with an alternate superblock:
e2fsck -b 8193 <device>

empeg:/empeg/bin# fsck -fay /dev/hda4
e2fsck 1.29 (24-Sep-2002)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/hda4: 7139/152576 files (7.8% non-contiguous), 3627572/4862970 blocks
empeg:/empeg/bin# swapoff /swapfile
swapoff: /swapfile: Invalid argument
empeg:/empeg/bin# sync
empeg:/empeg/bin#

I have recently converted my file system to EXT-3. Could this be why this no longer works? Is there some other way to correct this on a EXT3 FS partition? Everything still seems to be working fine but I'm nervous.

Thanks

I have recently converted my file system to EXT-3. Could this be why this no longer works?

Yes. I recommend you convert back to the stock file system.

you have to run fsck -fay /dev/hda5 instead.

Thank You!

In reply to:

---

Yes. I recommend you convert back to the stock file system.

---

Jeez Tony, what do you have against ext3? You participated in the previous thread that discussed this exact issue. Why not just link the poor guy to that thread or not say anything at all? People run plenty of non-standard stuff on their empegs and nobody expects you to cover it all in the faq or be the guy who can answer questions about everything, but this is the second thread where you come off as being openly hostile against ext3.

-Mike

FYI, this problem is now mentioned on my empeg ext3 page.

-Mike

I am not openly hostile against Ext3. But I think it's dangerous for someone to convert their player's filesystem if they don't know what they're doing. The person I responded to didn't sound like they were the appropriate customer for the Ext3 product, hence my recommendation to go back to the original filesystem.

I think that encouraging everyone to convert their filesystems before the stock software supports it is a dangerous thing to do. I think stiffer warnings should accompany your recommendations for it.

I think you need to edit the link you posted to read http://macgeek.dyndns.org/empeg/ext3/.

In reply to:

---

I think that encouraging everyone to convert their filesystems before the stock software supports it is a dangerous thing to do. I think stiffer warnings should accompany your recommendations for it.

---

Fair enough, and for the most part I agree with you. That is why most of the content on my info page is actually warnings about why a user may not want to use this software. There are benefits to running ext3 though and whether or not it is worth trying is the users decision. If the first response to anybody asking a question about it is always "go back to ext2" than nobody ever gets a chance to make up their own mind.

Also, I don't know how familiar you are with ext3, but it isn't really a "conversion" as much as an enhancement to ext2. The process of going back to ext2 is a one line command that does not result in data loss so if somebody does decide to try ext3 and then wants to go back to ext2 for whatever reason it isn't hard to do.

-Mike

In reply to:

---

I think you need to edit the link you posted to read http://macgeek.dyndns.org/empeg/ext3/

---

The bbs won't let me edit it anymore (it has been too long), but you are right, there is a typo in the link I posted and the one you posted works properly.

-Mike

Hmmm. I find your post interesting. At what stage of expertise do you feel a person is ready to take this stuff on? Personally I have learned more about Linux on my Empeg then I did on the box I set up with Mandrake. To me it is a perfect platform for learning the basics. What other systems could you make unrecoverable errors on and get the system back up and running in 10 minutes? For me it is not only a toy but a learning platform. I am not at risk of losing my music as it is backed up on both my PC and another Empeg. I think anyone who knows the risks and can rebuild their system is ready to experiment and learn. I've always found the members of this site to be very helpful and forthcoming with information. I have done the same with what little I know and as my knowledge increases I will give it back in what little ways I can.

Not to answer for Tony, but ....

There's a big difference between people who might want to learn Linux and people who might want their empeg to work ``better''. I think it's fair to assume the latter unless stated otherwise.

I think with ext3 isn't not a question of losing your music; If fact I think if you log into your player at all to run fsck yourself, you're qualified to use ext3.

A semantic change in fsck isn't really a sufficient reason to convert back, either.

Indeed, I agree, making the Empeg work better is the prime reason. The understanding of Linux and the improvement of the Empeg go hand in hand. You must possess a certain level of understanding of the operating system, specifically the one on the Empeg in order to use some of these upgrades. rsync is a good example. I'm tired of having to go through a process that takes 8 hours to sync my Empegs. Why not learn enough so I can use other methods such as rsync? EXT3 is so much nicer from my point of view. Would I rather wait for a minute and a half while the Empeg gets the hda4 into rw or would I rather not wait at all? The choice seems clear to me. Anyway, I'm not trying to start a debate here only find some answers and maybe someday I can help others that might find themselves in the same place.

Not to answer for Tony, but ....

Actually, you summed up my thoughts exactly.

My apologies, Neutrino, for incorrectly assuming that you were in "regular user" mode instead of "hacker" mode.

Hey, no hard feelings here Tony, you and the rest of the community have been awesome. I wonder if there is another open forum that contains so much brain power? I'm glad to be here.

I never really saw the point in ext3, and for this application the benefits are even less clear.

ext3 doesn't give any differences (improvements) over ext2 in read-only mode, which is the primary mode of the player. And it complicates the write path significantly while only giving doubtful benefits.

It creates more memory pressure, because it needs to keep data writes and meta-data updates 'ordered', not too helpful on a machine with only 12MB of memory. Especially when the ext3 extensions and the jbd layer have already added more unswappable kernel memory to the kernel image.

In data=ordered journalling mode it increases the amount of head seeks significantly because it has to sync all the data blocks before it writes the metadata updates to the journal, and then still has the update the actual ext2 inodes and bitmaps.

Data journalled does reduce seeks but it ends up being even worse, especially on an empeg, as it writes the data to both the journal and it's final location. And the bottleneck on the empeg is the relatively slow cpu->disk path which uses PIO instead of DMA.

The third mode that ext3 provides is useless as it only guarantees metadata consistency. It actually hides problems and lost data that fsck could normally detect.

And all of this to save an occasional fsck that in most cases isn't even needed and that in all the years of using Linux has only failed on me a few times. And that was because I was running a development kernel with buggy drivers that were writing blocks to the wrong location on disk, definitely not fsck or ext2's fault.

I guess I just don't get it.

Fscking 30gb when I crash my laptop playing in the kernel sucks hard?

I talked to Ted T'so about my concerns before enabling ext3, and he settled them; That was good enough for me, and since then, my laptop has been all good.

In fairness, you're talking about the empeg. So my laptop doesn't matter.

I will say that if you plan to crash the empeg while you're writing to it (well, not plan, but still...) it's nice to not worry about fsck.

In reply to:

---

I guess I just don't get it.

---

Interesting. As somebody who is doing active development on the empeg I guess I would expect you to "get it". As you mentioned there is no advantage to ext3 for read only operations. It is those times when it needs to be mounted rw that makes ext3 worth it for me. I spend a fair amount of time playing with third party software on my empeg and I am not that great about remembering to remount the drives read only. When I forget the 30+ minute fsck hurts. The thing that originally prompted me to get ext3 working was a couple of crashes while running a more unstable version of jemplode which resulted in long drive checks.

Ext3 could also be useful for applications that need to be able to store settings while the empeg is running. Gpsapp for example doesn't currently remember any of your settings between restarts. I assume at some point you will want to add the ability to save your waypoints, popup settings, etc. The easiest way to do that without using precious flash would be to temporarily remount the programs partition rw and write out a config file which would be a lot safer if the partition is journaled.

Anyway, those are a couple of the things that can be beneficial about journaling on the empeg. In my experience the potential performance issues you brought up haven't been an issue. I can still up and download files from the empeg at around 900k/sec. The memory issues could be a problem on the mk1, but on the mk2 and 2a there is a little extra ram to burn.

-Mike

I assume at some point you will want to add the ability to save your waypoints, popup settings, etc.

Even if so,

emporarily remount the programs partition rw and write out a config file which would be a lot safer if the partition is journaled.

It's not worth it just for this. It's not dangerous enough.

In reply to:

---

It's not worth it just for this. It's not dangerous enough.

---

Take it a step farther then. Kim Salo's nav software used its own swapfiles since the player didn't have enough ram for plotting larger routes. He also allowed you to record logfiles for debug purposes. Both of these required a partition to be mounted rw frequently.

-Mike

Still, if the partition is the programs partition, if I have to fsck it, it takes less time than loading my too-large database. Still not registering on the "I care" scale.

Take it a step farther then. Kim Salo's nav software used its own swapfiles since the player didn't have enough ram for plotting larger routes. He also allowed you to record logfiles for debug purposes. Both of these required a partition to be mounted rw frequently.


Yeah, that to me is going to be the big win for ext3 on the empeg. Jan makes some excellent points about ext2 - the reality is that ext2 is so bullet-proof in itself that the fscks are almost worthless - and if the journalling means we can remove them completely then it's almost worth doing just for the resulting end-user simplicity.

It's amusing how Micro$oft's scandisk.exe has left people in a state of fear when it comes to disk checks...I swear it's the first true example of cyber-terrorism.

Afaik, even with ext3 you cannot completely get rid of ext2. e2fsck doesn't just recover from lost updates during crashes or unclean shutdown (software problems). It can be used to validate/correct the filesystem integrity when the hardware is the cause of bad data, noise introduced by EMI while writing, tracks losing their data due to age/vibration/temperature.

empeg:/empeg/bin# fsck -fay /
e2fsck 1.29 (24-Sep-2002)
fsck: Is a directory while trying to open /

Just adding a footnote to an old thread...

tman pointed out a small problem in my ext3 setup instructions that actually causes this. The stock /bin/fsck does not need to be replaced, just /bin/fsck.ext2. If you leave the stock fsck it will call the newer fsck.ext2 which knows how to deal with ext3 and it works with the 'fsck -fay' syntax. I have updated my ext3 instructions so that new ext3 installs will be FAQ compatible. If you previously installed ext3 according to my instructions you can move '/bin/fsck.bak' back to '/bin/fsck' if you like.

-Mike