Re: sockets and HAMMER

2008-07-06 Thread Erik Wikström
On 2008-07-05 23:54, Matthew Dillon wrote: :Well, internal hard disks under 100G are getting hard to find, so I doubt :this is going to be a huge issue. It should be documented. : :There's also the possibility that things could be mounted under /, instead :of multiple partitions. Yah.

Re: sockets and HAMMER

2008-07-06 Thread Matthew Dillon
:Question: Is there a minimum recommended size of a PFS, or could you :create one per user? While it is not a complete replacement for quota it :would at least solve some of the same problems. : :-- :Erik Wikström You can't specify a size limitation for PFS yet so its kinda a moot

sockets and HAMMER

2008-07-05 Thread Gergo Szakal
I have noticed a strange problem on HAMMER: when UNIX domain sockets are created on my HAMMER partition, they cannot be accessed ('Socket operation on non-socket'), when I try to remove them, I get a 'Bad file descriptor' message. After a reboot they disappear. What is the reason for this? --

Re: sockets and HAMMER

2008-07-05 Thread Rumko
Gergo Szakal wrote: I have noticed a strange problem on HAMMER: when UNIX domain sockets are created on my HAMMER partition, they cannot be accessed ('Socket operation on non-socket'), when I try to remove them, I get a 'Bad file descriptor' message. After a reboot they disappear. What is the

Re: sockets and HAMMER

2008-07-05 Thread Gergo Szakal
On Sat, 05 Jul 2008 16:18:39 +0200 Rumko [EMAIL PROTECTED] wrote: Hmm, I also encountered 'Bad file descriptor' message (with a ls -la couldn't see the socket, but csh's autocomplete did find it and that message was displayed whenever I tried to rm it), but thought I messed up the machine one

Re: sockets and HAMMER

2008-07-05 Thread Matthew Dillon
:Gergo Szakal wrote: : : I have noticed a strange problem on HAMMER: when UNIX domain sockets : are created on my HAMMER partition, they cannot be accessed ('Socket : operation on non-socket'), when I try to remove them, I get a 'Bad file : descriptor' message. After a reboot they disappear. What

Re: sockets and HAMMER

2008-07-05 Thread Rumko
Gergo Szakal wrote: On Sat, 05 Jul 2008 16:18:39 +0200 Rumko [EMAIL PROTECTED] wrote: Hmm, I also encountered 'Bad file descriptor' message (with a ls -la couldn't see the socket, but csh's autocomplete did find it and that message was displayed whenever I tried to rm it), but thought I

Re: sockets and HAMMER

2008-07-05 Thread Gergo Szakal
On Sat, 05 Jul 2008 19:50:47 +0200 Rumko [EMAIL PROTECTED] wrote: Well, AFAIK HAMMER is still not production material, so it's just a matter of time before it can be used (since from what I read Matt intends to make HAMMER usable till the next release, it should be usable for anything but /

Re: sockets and HAMMER

2008-07-05 Thread Matthew Dillon
: Well, AFAIK HAMMER is still not production material, so it's just a : matter of time before it can be used (since from what I read Matt : intends to make HAMMER usable till the next release, it should be : usable for anything but / when 2.0 comes out?). : :It's obvious that bugs can occur,

Re: sockets and HAMMER

2008-07-05 Thread Matthew Dillon
The main issue for /tmp/ and /var is going to be size. HAMMER is not really designed for tiny partitions. They would have to be mounted 'nohistory' at a minimum and probably also have to be reblocked daily, depending on how full they get. The absolute bare minimum would be

Re: sockets and HAMMER

2008-07-05 Thread Justin C. Sherrill
On Sat, July 5, 2008 2:27 pm, Matthew Dillon wrote: The main issue for /tmp/ and /var is going to be size. HAMMER is not really designed for tiny partitions. They would have to be mounted 'nohistory' at a minimum and probably also have to be reblocked daily, depending on how

Re: sockets and HAMMER

2008-07-05 Thread Matthew Dillon
:Well, internal hard disks under 100G are getting hard to find, so I doubt :this is going to be a huge issue. It should be documented. : :There's also the possibility that things could be mounted under /, instead :of multiple partitions. Yah. I think what we are moving towards is more of a