I have 3 Linux distributions on my machine (Ubuntu, Arch and Fedora). I allocated half the RAM size (3GB) as swap for each of them. (I have realized that having 3 distributions is not of much use to me but I still have them.) The problem is now the 3 distributions mount all the 3 swap partitions ...
it's asking how NOT to share a swap partition. The other Q it's tagged as dupe of is asking if there are any side-effects of sharing swap. quite different questions.
@CraigSanders - yeah the problem is that it was poorly written so it was difficult to understand, as the comments indicate as well. In re-reading it and your comment it now makes a lot more sense, also Michael's answer help shed more light on it too. It still seems kinda sorta like a duplicate to me but you can twist my arm and I'll vote to reopen it if you concur.
Seems like the negative of a solution is a duplicate especially when the positive includes details on how to do the negative?
@CraigSanders - let me know if you want to reopen!
@Gilles I found this page: subversion.apache.org/docs/community-guide/debugging.html. But it looks like you have to have this enabled via compilaiton (--enable-maintainer-mode). It enables a SVN_DBG env. var. Look through that page, further on they show debugging with a proxy and wireshark
@slm, yeah i'd like to re-open. i made three quite long comments Michael's answer that i'd prefer to re-write as a supplemental answer to his, and add more detail/examples than is possible in a comment.
Sorry about closing that, when I read it it completely sounded like the opposite, I've edited the Q to try and bring out the fact that it's the opposite
@derobert - assuming you saw my Q: unix.stackexchange.com/questions/91792/…. Getting sick of seeing them on the site and they're all slightly differnt forms of the same one
@derobert - go for it, I kind of wanted to collaborate on that answer rather than attempt to make one myself, I know a decent amount about ntp, but there are better ways to do things that I'm not that confident on.
Yes, I don't think there is (for example & for at least the last 4+ years) a good reason to have both ntpd and ntpdate. Not since ntpd gained -g, at least.
@derobert - yeah go for it with an answer, understand if you want to retain control of it, if there is any points missing I/others can create complimentary answers to the Q, or feed the info to you for reviewing prior to addition. I just don't want the Q to become one with 5+ answers, I'd like it to have 1-2 max.
but lay the foundation out in easy to follow steps to good NTP practices vs.bad, seems like a long time coming for that topic, no?
ultimately I was hoping to create a canonical answer and then pull those other Q&A into it so we could start to limit there damage of mis-info going forward
I figure an answer needs, at minimum, (1) quick discussion of timekeeping on PC hardware (2) quick explanation of what ntpd does, in theory; (3) config that works on sane/good hardware; (4) a bunch of ways to make it work on broken hardware
One thing which (from my quick skim) it seems that Stephane didn't mention is that which has no idea about your shell's path hash table. This has the effect that it might return a result which is not representative of what actually is run, which makes it ineffective in debugging.
I guess I'll try and write something up Tuesday, or today if I find some time... Going to take some research, to make sure I get it right. And also to dig up links to all the various bits of ntp docs, to give for further info
yeah on Q's that have 4-5 A's already it's rare that there is anything extra to add, ppl just looking to try and get started with some rep is usually the case
it's only has 39 views so the Q is having the effect of not a lot of attempters at answering it
i told gilles & terdon this was my way of forcing myself to answer it once in for all by asking the Q. I typically answer my own Q's to try and fill in missing domain knowledge on the site
but if a resident expert can do better I'm all for moving out of the way 8-)
ultimately i just wanted the resource to reference when we get the Q's later on about it so the info isn't so fractured and incorrect
@derobert I'd suggest to add something like "VMs with saved states and shapshots" to the weird section. I remember that there was a question regarding that not long ago, and I also have that issue myself all the time
@MartinvonWittich - if you roll the edits back on the Q you'll see we took out references to VM to try and focus the Q. The A can definitely include that info but were trying to keep the Q from becoming too broad.
One thing which (from my quick skim) it seems that Stephane didn't mention is that which has no idea about your shell's path hash table. This has the effect that it might return a result which is not representative of what actually is run, which makes it ineffective in debugging.
It's still one of my fav. Q&As more b/c of how much knowledge Stephane possess though, the A is so long and exhaustive that i can never fully internalize even 1/2 of it
@derobert zsh has amazing features. I really like its argument completion for example: martin@dogmeat ~ % ls -- option --all -- list entries starting with . --almost-all -- list all except . and .. --author -- print the author of each file --block-size -- specify block size [...]
it doesn't support all commands of course, but it's very helpful for that commands that it does support.
you have to enable bash-completion to get it, I assume zsh has something similar (e.g., some set of scripts to tell it about various commands, and their options)
@derobert about ntpd, I don't know if this noteworthy, but it had a bug some time ago that it stopped working when your internet connection is a PPP interface and this interface got a new IP address. If that is still the case, you could mention it under "when it doesn't work" - the fix is to add /etc/init.d/ntpd restart to /etc/ppp/ip-up.d/somefile
@derobert nevermind, I've found bug reports... seems like this was fixed in 4.2.2 http://bugs.ntp.org/show_bug.cgi?id=51 http://bugs.ntp.org/show_bug.cgi?id=622 Even Debian squeeze has 4.2.6, so it's probably not noteworthy anymore
Yeah bash added the complete facility which is customizable, you can add completions for anything
@MichaelMrozek - yeah I use the remote all the time, it can look inside your ssh .config file and complete on those too
@MichaelMrozek - i maintain a variable type of .config file so that i can go to servers over a proxy by adding a -o to their name (outside) vs. without (defaults to the normal name of the host)
@MichaelMrozek - example: ssh skinner-o vs. skinner. The first will go over a proxy like this: "ProxyCommand ssh sam@external.host.com nc echo %h|sed 's/-o//' %p"
it can do the remote files on the server too, so this is extremely powerful for reaching into another computers file system and hitting tab completions to see the files without having to ssh into the server and poke around first
@derobert bash does remote completion on scp? How? Is there a flag somewhere?
Ah, I see, works on my LAN, does not seem to on WAN
@slm I have a (passwordless) shh over a non-default port set up and bash won't complete the remote path, it tries and hangs for ever (which is why I thought it dd not work) works fine on local though. Strange
Damn, I can't believe I did not know this :) Great feture
@terdon have you put the necessary configuration into your .ssh/config? If you use the hostname that you have configured in the config, bash should be able to figure out the correct path from there
Also a very nice thing for remote completion: if you want to use it with servers that expect a password, configure ssh controlmaster. You can then open a connection to the server in one window, and use remote completion in all other windows because ssh will then use the existing connection. It's also faster because it doesn't need to open new connections all the time.
@terdon because bash is parsing that line, and it probably fails to realize that it has to use port 24222. It probably just looks at the terdon@138.124.66.43:/cobeli part of the command line
if you're working in multiple windows with interactive sessions, and you try to close the first one (CTRL+D or exit), the ssh session will hang and remain open. If you then kill it with CTRL+C, all other sessions will close too.
Not that I would know of. At least remote completion to passwordless servers will be noticeably faster if you already have a session to that server, because the new ssh sessions will spawn instantly
with open session: martin@dogmeat ~ % time ssh dev2 true 0:00.02 ssh dev2 true
@terdon, yes controlmaster is your friend, just remember that if you have like 3 connections they're sharing one pipe, so if you take the original one down (i.e. exit it) the others go with it!
yeah that's a extremely useful productivity trick if you ssh to a lot of machines
i used to work at kodak and would use that to reach into the data centers since they were islands that could only be accessed from specific host IPs for security reasons
@slm what's even nicer: they can be real host names and then they will override the hostname. For example, if you have a server example.com that listens on port 2222 instead of 22, you can just put this in your config: Host example.com Port 2222 and ssh will handle the different port transparently.
We have about 800 customer servers, and some of them have been configured for different ports and such. We have a field for that in our CRM where we enter the deviating hostnames/port and stuff, and then a .ssh/config is autogenerated from this configuration. That way we don't have to memorize it all, we just ssh to the domain and the config will take care of the rest.
:) Would be if I could claim to understand everything that does :) Managed to set it up and it works to tunnel to my VM through another host is all I know
Cool, thanks. Don't really need much though, all I do is connect to my lab which has to be done through a gateway and then connect o one of 4 servers that I have to play with.
How can I change a owner of a file in AIX? I created a file using my personal account and then I tried to change user for this file:
chown rootuser myfile.csv
chown: myfile.csv: Operation not permitted.
Then I changed to root user
sudo /usr/bin/su - rootuser
and tried to change the owner o...
When you're using ext4, you can check for badblocks with the command e2fsck -c /dev/sda1 or whatever. This will "blacklist" the blocks by adding them to the bad block inode.
Sorry, but that's complete nonsense. e2fsck won't check for bad blocks on the physical hard disk and it also won't "bl...
That's like the tenth time I've explained how to use smartctl to run a selftest, and I'm only active for like two weeks
One of these days, I'm gonna write a detailed howto and just link to that...
@terdon no, "sudo su" is completely pointless. These people all need to learn about sudo -i or sudo -s ._.
-c This option causes e2fsck to use badblocks(8) program to do a read-only scan of the device in order to find any bad blocks. If any bad blocks are found, they are added to the bad block inode to prevent them from being allocated to a file or direc‐ tory. If this option is specified twice, then the bad block scan will be done using a non-destructive read-write test.
... RTFM :-P
@MartinvonWittich as you can see from answer, I agree with you that the hard disk really ought to map out the sector itself
I'm pretty sure that we do have a question on how to actually get a SATA (etc.) drive to remap a sector
@MartinvonWittich BTW: At least you haven't had to write a bunch of answers yelling at people to please stop trying to destroy their data with mdadm :-P
The tl;dr: how would I go about fixing a bad block on 1 disk in a RAID1 array?
But please read this whole thing for what I've tried already and possible errors in my methods. I've tried to be as detailed as possible, and I'm really hoping for some feedback
This is my situation: I have two 2TB d...
@derobert well, when I see "read failure" after a selftest, than that's it for me. A hard disk is not expensive compared to downtime or even data loss.
Also isn't a hard disk supposed to use reserve blocks as long as it can, and only report a failure when there are no more reserve blocks left?
@MartinvonWittich yes, depends on the disk. Sometimes the disk will reallocate. Other times the disk will attempt a write to the block, read it back, and say "hmm, OK now, let's keep using this sector"
One thing which (from my quick skim) it seems that Stephane didn't mention is that which has no idea about your shell's path hash table. This has the effect that it might return a result which is not representative of what actually is run, which makes it ineffective in debugging.
How can I change a owner of a file in AIX? I created a file using my personal account and then I tried to change user for this file:
chown rootuser myfile.csv
chown: myfile.csv: Operation not permitted.
Then I changed to root user
sudo /usr/bin/su - rootuser
and tried to change the owner o...
When you're using ext4, you can check for badblocks with the command e2fsck -c /dev/sda1 or whatever. This will "blacklist" the blocks by adding them to the bad block inode.
Sorry, but that's complete nonsense. e2fsck won't check for bad blocks on the physical hard disk and it also won't "bl...
The tl;dr: how would I go about fixing a bad block on 1 disk in a RAID1 array?
But please read this whole thing for what I've tried already and possible errors in my methods. I've tried to be as detailed as possible, and I'm really hoping for some feedback
This is my situation: I have two 2TB d...
