Re: [zones-discuss] Code review for zones support in SNAP (updated)
Joe, Thanks. We'll make that correction. -ethan Joseph J VLcek wrote: > Ethan Quach wrote: >> Joe, Jean, Jerry, >> >> >> Thanks for providing your comments. We've built an incremental webrev >> with just the changes from your comments. This webrev also includes a >> couple of dev bugfixes that went in during the code review process - >> 3776, 3772 >> Any further comments are welcomed. >> >> http://cr.opensolaris.org/~equach/webrev.snap_zones.2 >> >> >> As a reference, we've also refreshed the overall webrev here: >> >> http://cr.opensolaris.org/~equach/webrev.3686 >> >> >> Thanks again for the review, >> >> -ethan > > > This all looks good. I only found one small nit. > > Joe > > >usr/src/lib/libbe/be_create.c > +-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+ > > > Issue: > -- > > Typo-s in comment: > > Change from: > 1004 * root dataset of the new BE. The uuid is use to set the parentbe > 1005 * property for the new zones datasets. > To: > 1004 * root dataset of the new BE. The uuid is used to set the parent be > 1005 * property for the new zones datasets. ___ zones-discuss mailing list zones-discuss@opensolaris.org
Re: [zones-discuss] Code review for zones support in SNAP (updated)
Ethan Quach wrote: > Joe, Jean, Jerry, > > > Thanks for providing your comments. We've built an incremental webrev > with just the changes from your comments. This webrev also includes a > couple of dev bugfixes that went in during the code review process - > 3776, 3772 > Any further comments are welcomed. > > http://cr.opensolaris.org/~equach/webrev.snap_zones.2 > > > As a reference, we've also refreshed the overall webrev here: > > http://cr.opensolaris.org/~equach/webrev.3686 > > > Thanks again for the review, > > -ethan This all looks good. I only found one small nit. Joe usr/src/lib/libbe/be_create.c +-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+-=+ Issue: -- Typo-s in comment: Change from: 1004 * root dataset of the new BE. The uuid is use to set the parentbe 1005 * property for the new zones datasets. To: 1004 * root dataset of the new BE. The uuid is used to set the parent be 1005 * property for the new zones datasets. ___ zones-discuss mailing list zones-discuss@opensolaris.org
Re: [zones-discuss] Code review for zones support in SNAP
Ethan Quach wrote:
>
> Comments, mostly on libbe ...
>
>
> General comment - Since we don't yet support boot environments
> within a zone, beadm should fail gracefully when run in a zone, like:
>
># beadm
>beadm is not supported in a zone.
>
> rather then trying to do something and failing miserably. Most of
> the if(GLOBAL_ZONEID) checks we have in libbe now are really
> just primers for when we build in that support.
>
>
> usr/src/man/beadm.1m.txt
> --
> 333 - SUNWinstall -> SUNWbeadm
>
>
> messages.py
>
> 126 - BE_ERR_NO_ACTIVE_ROOT should be in here.
>
>
> be_activate.c
> -
> 136 - initialize to { 0 }
fixed
>
> 169-199 - This whole chunk needs to be encapsulated inside a
>if (GLOBAL) {
>
>}
moved
>
> 215-221, 241-245 - These two chunks seem like they could be
> moved up into the if (GLOBAL) {} from 169.
moved
>
> With some refactoring, the zfs_promote() calls in set_canmount()
> could probably be ripped out and replaced by calls to be_promote_zone_ds()
> (renamed to something more generic like be_promote_ds() of course).
> I never understood why the zfs_promote() calls were hidden in the
> set_canmount() function anyway. If we don't get to this per the code
> review changes, please file a bug to track this.
refactored this so that the promote's are now done with calls to
be_promote_ds_callback and are no longer in set_callback (where they never
really should have been in the first place...).
>
> 972-1023 - This whole chunk can be removed if you replace 1025-1026
> with a direct call to the callback giving it the root dataset of the
> zone BE.
>
>i.e. replace 1025-1026 with:
>
>if (be_promote_sub_ds_callback(z_zhp, NULL) != 0) {
done
>
> 1028 - remove "subordinate"
>
> Rename the following functions to be more generic:
>be_promote_zone_ds() --> be_promote_ds()
>be_promote_sub_ds_callback() --> be_promote_ds_callback()
I renamed be_promote_sub_ds_callback() to be_promote_ds_callback() However I
left be_promote_zone_ds() as is since it's doing a bunch of zones lookup to
find
the zones and the correct zones datasets to promote before calling
be_promote_ds_callback().
>
>
>
> be_create.c:
> ---
> 1880 - future
fixed
>
> 1878-1883 - this comment needs to be removed. It doesn't apply to
> this section of code.
fixed
>
> 1884 - within the context of this function - that is, code running in the
> global zone that's processing non-global zones - this if() can be removed,
> along with 1895.
I removed them.
>
> 1944 - A comment stating this is getting the uuid of the newly cloned
> parent BE would be nice.
comment added
>
> 2772-2774 - Missing function block comment.
>
>
> libbe.h:
> -
> 117 - "not" -> "no"
> 177 - Remove this line; its not used anywhere.
>
>
>
> thanks,
> -ethan
>
>
>
> Evan Layton wrote:
>> Hello All,
>>
>> We're down to the wire on the zone support changes to SNAP upgrade and
>> are looking for code review comments. We'll be taking comments up
>> until COB Tuesday October 7th. Your comments are as always welcome and
>> appreciated.
>>
>> Defect 3686 is the blocker bug that was submitted to cover this work
>> and the webrev is available at:
>>
>> http://cr.opensolaris.org/~equach/webrev.snap_zones/
>>
>> Thank you in advance for your comments and help!
>> -evan
>> ___
>> zones-discuss mailing list
>> zones-discuss@opensolaris.org
>>
___
zones-discuss mailing list
zones-discuss@opensolaris.org
Re: [zones-discuss] Code review for zones support in SNAP
Evan Layton wrote: > Hello All, > > We're down to the wire on the zone support changes to SNAP upgrade and are > looking for code review comments. We'll be taking comments up until COB > Tuesday > October 7th. Your comments are as always welcome and appreciated. > > Defect 3686 is the blocker bug that was submitted to cover this work and the > webrev is available at: > > http://cr.opensolaris.org/~equach/webrev.snap_zones/ > > Thank you in advance for your comments and help! > -evan > ___ > zones-discuss mailing list > zones-discuss@opensolaris.org usr/src/lib/libbe/be_create.c: line 1306 - the goto isn't needed line 1381 - We're not collecting any errors that may have been returned from be_get_snap and just returning a generic BE_ERR_ZFS be_errno. Can we collect any useful information from errors returned by this call? line 1420 - We're not capturing any errors here and just returning whatever "ret" was set to previously, which in this case appears to be 0. line 1519 - We should be collecting the error returned from _be_mount instead of returning the generic mount error. line 1539 - same as 1519 but for unmount. line 1635 - Why not just "goto done" instead of the ZFS_CLOSE and return? unless the callback function has already closed the zfs handle. line 1638 - shouldn't the callback function be_destroy_zone_roots_callback have already closed the zfs handle? line 1705 - shouldn't this callback function always close the zfs handle? line 1719 - if the comparison fails we don't close the zfs handle here. line 2772 - We need a comment header for this function. moving on to be_mount.c -evan ___ zones-discuss mailing list zones-discuss@opensolaris.org
Re: [zones-discuss] Code review for zones support in SNAP
Comments, mostly on libbe ...
General comment - Since we don't yet support boot environments
within a zone, beadm should fail gracefully when run in a zone, like:
# beadm
beadm is not supported in a zone.
rather then trying to do something and failing miserably. Most of
the if(GLOBAL_ZONEID) checks we have in libbe now are really
just primers for when we build in that support.
usr/src/man/beadm.1m.txt
--
333 - SUNWinstall -> SUNWbeadm
messages.py
126 - BE_ERR_NO_ACTIVE_ROOT should be in here.
be_activate.c
-
136 - initialize to { 0 }
169-199 - This whole chunk needs to be encapsulated inside a
if (GLOBAL) {
}
215-221, 241-245 - These two chunks seem like they could be
moved up into the if (GLOBAL) {} from 169.
With some refactoring, the zfs_promote() calls in set_canmount()
could probably be ripped out and replaced by calls to be_promote_zone_ds()
(renamed to something more generic like be_promote_ds() of course).
I never understood why the zfs_promote() calls were hidden in the
set_canmount() function anyway. If we don't get to this per the code
review changes, please file a bug to track this.
972-1023 - This whole chunk can be removed if you replace 1025-1026
with a direct call to the callback giving it the root dataset of the
zone BE.
i.e. replace 1025-1026 with:
if (be_promote_sub_ds_callback(z_zhp, NULL) != 0) {
1028 - remove "subordinate"
Rename the following functions to be more generic:
be_promote_zone_ds() --> be_promote_ds()
be_promote_sub_ds_callback() --> be_promote_ds_callback()
be_create.c:
---
1880 - future
1878-1883 - this comment needs to be removed. It doesn't apply to
this section of code.
1884 - within the context of this function - that is, code running in the
global zone that's processing non-global zones - this if() can be removed,
along with 1895.
1944 - A comment stating this is getting the uuid of the newly cloned
parent BE would be nice.
2772-2774 - Missing function block comment.
libbe.h:
-
117 - "not" -> "no"
177 - Remove this line; its not used anywhere.
thanks,
-ethan
Evan Layton wrote:
> Hello All,
>
> We're down to the wire on the zone support changes to SNAP upgrade and are
> looking for code review comments. We'll be taking comments up until COB
> Tuesday
> October 7th. Your comments are as always welcome and appreciated.
>
> Defect 3686 is the blocker bug that was submitted to cover this work and the
> webrev is available at:
>
> http://cr.opensolaris.org/~equach/webrev.snap_zones/
>
> Thank you in advance for your comments and help!
> -evan
> ___
> zones-discuss mailing list
> zones-discuss@opensolaris.org
>
___
zones-discuss mailing list
zones-discuss@opensolaris.org
