gjournal: journaled slices vs. journaled partitions

[Volodymyr Kostyrko]

Carl wrote:
> My goal is to build a 2-disk server configured with gmirror and gjournal 
> for maximum reliability. There will never be a second operating system 
> on the system, but I prefer not to freak out any non-FreeBSD repair 
> tools that might be used, so I will use compatibility instead of 
> dangerously dedicated mode. This means I need one slice, but see no 
> reason for more. Inside that one slice will be the usual array of 
> partitions (ie. /, swap, /var, /tmp, /usr, /data).
> 
> Now, I think gmirror allows me to mirror the entire drive rather than 
> forcing me to do per-slice or even per-partition mirroring. I'm looking 
> for the simplest in-field replacement procedure when one of the drives 
> dies and I imagine a whole drive mirror achieves this. Am I right?
> 
> gjournal, OTOH, has me really confused. The man page for gjournal(8) 
> specifically does not recommend that small partitions be journaled. I 
> assume that's because the journal provider rivals the partition in size 
> and is therefore overhead heavy. It seems to me, though, that if I can 
> journal the slice as a whole instead of per-partition journaling, that 
> there will essentially then be only one journal provider for the 
> combination of all partitions (ie. slice) and that the aforementioned 
> overhead becomes minor. Having smaller partitions included in journaling 
>  seems like a good thing to me. So how do I achieve per-slice journaling 
> instead of per-partition? Every time I read up on someone else's 
> gjournal implementation, it seems to end with adding <partition>.journal 
> entries to /etc/fstab. Am I trying to achieve the impossible or 
> ill-advised here?

I have some setups were gjournal was put on device rather the on 
partition, i.e.:

[umgah] ~> gmirror status
          Name    Status  Components
mirror/umgah0  COMPLETE  ad0
                          ad1
[umgah] ~> gjournal status
                  Name  Status  Components
mirror/umgah0.journal     N/A  mirror/umgah0
[umgah] ~> glabel status
             Name  Status  Components
   ufs/umgah0root     N/A  mirror/umgah0.journala
label/umgah0swap     N/A  mirror/umgah0.journalb
    ufs/umgah0usr     N/A  mirror/umgah0.journald
    ufs/umgah0var     N/A  mirror/umgah0.journale
[umgah] ~> mount
/dev/ufs/umgah0root on / (ufs, asynchronous, local, noatime, gjournal)
devfs on /dev (devfs, local)
/dev/md0 on /tmp (ufs, asynchronous, local)
/dev/ufs/umgah0var on /var (ufs, asynchronous, local, noatime, gjournal)
/dev/ufs/umgah0usr on /usr (ufs, asynchronous, local, noatime, gjournal)
devfs on /var/named/dev (devfs, local)

And yes, mirror autosynchronization is turned off, gjournal takes care 
of that too.

It's not stated in manual, but gjournal is typically transparent for any 
type of access, just in case of UFS file system is marked as journaled 
so any metadata writes can be distinguished from data writes. Without 
that gjournal does literally nothing.

-- 
Sphinx of black quartz judge my vow.