On 01/07/2015 03:10 AM, Ric Moore wrote: > On 01/06/2015 07:23 PM, ~Stack~ wrote: > >> I keep seeing all of these posts online saying how easy it is to disable >> systemd from runing fsck because it "honors" the '0' in the sixth field >> of /etc/fstab. Well that's just pure bull$h1t... That was one of the >> first things I tried some time ago. As far as I can tell on neither of >> my Jessie machines (one physical one virtual) does systemd honor the >> fstab in terms of doing a fsck. All of the partitions are set to 0 in >> /etc/fstab. > > That doesn't make it "bullshit", it means that in your instance it > doesn't help. Instead, just maybe your system is trying to tell you > something when it continually forces fsck. Read the man page. You are right in that there is something probably wrong with my instance and it might be trying to tell me something. However, I *have* read the man page and the man page for fstab says "If the sixth field is not present or zero, a value of zero is returned and fsck will assume that the filesystem does not need to be checked." Therefore, I am explicitly telling systemd-fsck to buggeroff and not check. I as the owner of the device have acknowledged and assumed all the risk for not running fsck on my drive. The fact that it is /still/ doing it against my wishes is what annoys me. > The trick is to get your poor stupid dumb machine to tell it's human > where it hurts and how to fix it. It's like dealing with a puppy that > whines. I will not deny that this laptop is more than a little weird. It's almost 10 years old and has seen its share of hardship in life. But as a secondary laptop I can haul around with me and not worry about it, it's a fantastic system. > > OK, your next message reveals that you are encrypting your drives, > including swap?? Insert "encrypted hard drive fsck" into your search > bar. Lately, others report problems. I did look for that. Obviously not in the right places, but even typing that exact string into my search bar returns results for people who have errors to work with and search against. I don't. All my messages say that fsck returned clean with the exception of the one that goes by too fast and says something about a timeout. > Another thought would be to disable swap, if you have enough memory to > see if that helps at all. man swapoff Well, since I figured out that systemd-fsck goes really quickly when I have /dev/sda3 instead of the UUID for the swap partition, it works just fine. I don't see how turning off swap all together will suddenly make systemd-fsck start listening to the boot parameter and fstab telling it not to run fsck at all. > > Since it was encrypted no telling what this would do: man swapoff > -------------------------------- > swapoff-f, --fixpgsz > Reinitialize (exec /sbin/mkswap) the swap space if its page size > does not match that of the current running kernel. mkswap(2) > initializes the whole device and does not check for bad blocks. > ------------------------------- # swapoff -a # swapon --fixpgsz /dev/dm-1 # swapon -a # swapon -s Filename Type Size Used Priority /dev/sda3 partition 585724 0 -1 No errors or anything, so I assume that is good... > At any rate all the fsck'ing is telling you something is broken. That is > what it is supposed to do. Not when I have told systemd-fsck to stop trying. Then it is supposed to not do it. Also, I have yet to see a report that says something other than clean. Using the information I have, I am leaning toward that something broken being systemd-fsck...I have run out of ideas on how to tell it to stop... If you are running luks check this: > http://serverfault.com/tags/luks/hot > > Happy hunting troubleshooting with a shotgun. :) Ric Thanks! I do appreciate the help! ~Stack~
Attachment:
signature.asc
Description: OpenPGP digital signature