new user - quick test of dragonflybsd

[Joris Giovannangeli]

Hi

i've probably not the most useful/accurate answers, but here are some.

> 1) Why "wired" memory is so high - over 120MB. How to check what
> actually takes that much.
> 
> vmstat -m shows few megabytes, i stripped kernel binaries so kernel+few
> modules are less than 11MB. even looking at vmstat -s doesn't sum up.
> 
> Actually it is similar to FreeBSD, yet i still didn't get definite
> answer how to check what actually takes that memory. all data i was
> advised to check simply don't sum up.

120MB is really low. Wired memory is unpagable memory. I might be wrong
but afaik vmstat -m shows memory statistic for paged allocated by
kmalloc. It does not account for, for instance, page tables, which is
one of the main large source of wired memory.

There is a page table optimization which can further reduce this number
(sysctl vm.pmap_mmu_optimize), but it is not yet stable. It's a bit more
stable on -master but it's not stable at all on 3.6. Hopefully it will
be fixed and it might help here depending on your workload (if you have
a lot of shared memory mapping, for instance postgresql server)

> 2) I can mount whole hammer filesystem with "nohistory" option. Just as
> well as noatime that i do on all systems no matter what filesystem i use.
> 
> But how can i have nohistory mode on selected PFS?

I don't know how to do that. You can use chflags nohistory and
noshistory to control that at a per file basis. You can also set up the
pfs (with hammer viconfig) with a short snaphsot period and a short
shapshot retention time, and set up a short prune period. It won't be
the same as nohistory, but it will delete the history early. But you
won't reclaim space before actually running reblock.

My guess is that hammer is not designed to run like that.

> 3) can i set PFS up so all history points would be visible under some
> subdirectory?
> 
> I would use it for samba-exported PFS so user would be able to browse
> it, and eg. recover older versions of files without asking me for help.

I don't think so, but that's what snaphsots are for. You can set up a
pfs to take a snaphot every ten seconds for instance, and all of the
snaphots will be visible under the snaphot directory (which is
/var/hammer/<pfs-name> by default, but you can change this setting per pfs)

> 4) I'm getting once every few minutes under load messages like
> 
> "[diagnostics] had to recurse on directory /XXX"
> 
> what is it?
> 
> except this messages from kernel everything works OK.

no idea on this one.

> BTW ability to keep swapcache over reboots would be great. Otherwise -
> cache is empty when it is most needed - after reboot when there is high
> I/O rate.

It would be, indeed, but it's not possible due to the design of
swapcache. Swapcache uses the kernel live structures representing files,
which are not persistent across a reboot.

> 
> 
> Finally
> 
> 6) Why there is a limit of 65536 PFS's? No problem for me now, but on
> large server it would be useful to make one PFS per user and it may be a
> problem then.
> 
Dunno exactly why, but 65536 is 2^16, probably some limitation in the on
disk format somewhere.

Regards,
joris