Barbara.Lundquist at Sun.COM wrote: > I'll change back the titles to say create instead of clone, But, I'd
OK. > also like to add a new section title about > how to clone and include your comments below plus point them back to the Cloning is the method libbe uses to create a new BE so I'm not sure how relevant it is for this document since we are talking about managing BE's with beadm. Mentioning how libbe clones snapshots in one case and copies in another case doesn't seem pertinent either until we support creating BE's on a different pool. > How to Create procedures for specifics. > I'd like people to see "how to copy a BE" in the TOC, so they don't have Is "How to copy" any easier then "How to create" for customers? If not than "How to create" is better since there is no "beadm copy" subcommand and intermingling the two (copy & create) isn't as clear to me as just referring to create. Thanks Tim > to hunt for that info. > Barbara > > > On 10/24/08 12:47, Tim Knitter wrote: >> Barbara.Lundquist at Sun.COM wrote: >> >>> Isn't is useful for people to know how to copy or clone a boot >>> environment? We don't clarify in the command itself that the create >>> command actually does clones. I think we need to explain that in the >>> document. >>> >> >> I think we should, when talking about the "beadm create" subcommand, always >> state that we are creating a new BE from an existing BE. Under the covers we >> either clone or copy which doesn't need to be exposed in the doc at this >> time IMHO. Or if we do want to describe cloning/copying then we need to say >> that the beadm create subcommand clones the datasets that make up a BE when >> that new BE is in the same pool as the source BE. We copy when creating a >> new BE in a separate pool. >> >> Tim >> >> >>> Barbara >>> >>> >>> On 10/24/08 11:14, Tim Knitter wrote: >>> >>>> Hi Barbara - Here are my comments. >>>> >>>> p5: >>>> >>>> "A boot environment is an image that is bootable" >>>> >>>> I don't think using "image" is good to define what a boot environment is >>>> since there is no definition of what "image" is in the doc either. >>>> >>>> A boot environment is an installed, bootable instance of an operating >>>> system... >>>> >>>> mount or unmount -> mount, unmount, list >>>> >>>> Should probably add the list sub-command here as well. >>>> >>>> p6 bullet 1: >>>> >>>> "After successfully completing the changes" -> After successfully >>>> completing the >>>> update >>>> >>>> The above paragraph talks about updating so "changes" should be replaced >>>> with "update" >>>> >>>> p10: >>>> >>>> Note ? You must be root on your system to use the beadm utility. >>>> >>>> Nit: >>>> We could also mention that beadm is part of the "Software Installation" >>>> profile and a user who is allowed to be part of this profile can use >>>> pfexec to execute beadm. >>>> >>>> p11: >>>> >>>> 3 Review the information that is displayed. >>>> >>>> Is this to be filled in later? >>>> >>>> p12: >>>> >>>> "How to Clone a Boot Environment" >>>> >>>> Since we don't have a clone subcommand, and this section is talking about >>>> "Creating a Boot Environment", this could be rewritten to say: How to >>>> Create a New Boot Environment >>>> >>>> Using clone in the examples doesn't seem right to me even though a clone >>>> is what the resultant "beadm create" command creates. IMO all examples >>>> should remove clone and replace with creating a new Boot Environment to be >>>> consistent. >>>> >>>> p13: >>>> >>>> "Creating a New, Cloned Boot Environment Dataset (BE2)" >>>> >>>> This implies that we are creating a single dataset but the example shows >>>> many datasets. Please revise. Something like: Creating a New Boot >>>> Environment with Multiple Datasets (BE2) >>>> >>>> "The following example illustrates the dataset in a newly created boot >>>> environment." >>>> >>>> dataset -> datasets >>>> >>>> "The original dataset in this example is BE1." >>>> >>>> original dataset -> original Boot Environment >>>> >>>> "The new boot environment, BE2, is contained in its >>>> own dataset. The new boot environment can also contain separate datasets >>>> for traditional file >>>> systems, such as /var or /opt." >>>> >>>> -> >>>> >>>> The new boot environment, BE2, contains separate datasets cloned from BE1. >>>> If BE1 contains separate datasets for traditional file systems, such as >>>> /var or /opt, then those datasets are also cloned. >>>> >>>> Creating a Clone With Shared File Systems >>>> -> >>>> Creating a New Boot Environment With Shared File Systems >>>> >>>> The following example illustrates the dataset >>>> -> >>>> The following example illustrates the datasets >>>> >>>> p14: >>>> >>>> rpool/export/ remove trailing '/' >>>> >>>> How to Clone an Inactive Boot Environment >>>> -> >>>> How to Create a Boot Environment from an Inactive Boot Environment >>>> >>>> p15: >>>> >>>> $ beadm create beName at snapshot >>>> >>>> The example uses "snapshot" but the text talks about >>>> "snapshotdescription". Please use one or the other throughout this section. >>>> >>>> "If you have an older snapshot of an environment that you want to use as >>>> default in the GRUB >>>> menu, for example, you would copy that snapshot, making a new boot >>>> environment that can be activated, and then booted, from the GRUB menu." >>>> >>>> This paragraph seems confusing and is not needed for this example. Please >>>> remove. >>>> >>>> p16: >>>> >>>> "You can change an inactive boot environment into an active environment." >>>> >>>> active environment -> active boot environment >>>> >>>> General note... Part of the document talks about "upgrade" and other parts >>>> of it mention "update". Please be consistent throughout. "Upgrade" should >>>> be replaced with "update" and it should be clear that an update happens >>>> via "pkg image-update". >>>> >>>> p17: >>>> >>>> "The boot environment is mounted but remains inactive. You can update the >>>> packages this >>>> mounted, inactive boot environment." >>>> -> >>>> The boot environment is mounted but remains inactive. >>>> >>>> The next section talks about updating the BE so the last sentence should >>>> be removed or modified to mention the next section. >>>> >>>> p18: >>>> >>>> "Destroying Unneeded Boot Environments" >>>> >>>> Is "unneeded" really needed here? :-) It isn't mentioned in the text below >>>> it and seems unnecessary. >>>> >>>> General note... The doc talks about filesystems and datasets >>>> interchangeably. A filesystem is a type of dataset so it should be used >>>> sparingly and dataset used instead in most cases. e.g. >>>> >>>> "See the following example, where BE1 and BE2 share the rpool/export and >>>> rpool/export/home file systems. The datasets for BE1, BE2, and the shared >>>> file systems >>>> include the following: >>>> >>>> rpool/ROOT/BE1 >>>> rpool/ROOT/BE1/opt >>>> rpool/ROOT/BE2 >>>> rpool/ROOT/BE2/opt >>>> rpool/export >>>> rpool/export/home >>>> >>>> In this example all these can either be referred to as datasets or >>>> filesystems. Datasets being the more appropriate since we don't know what >>>> the backing store is and it is assumed to be a file system in your >>>> example. If we just mention "dataset" then there is no need to mention >>>> "filesystem" >>>> >>>> p19: >>>> >>>> " If you want to rename the active environment, first, make a different >>>> environment >>>> active." >>>> -> >>>> If you want to rename the active boot environment, first, make a >>>> different boot environment >>>> active. >>>> >>>> "If the new name already is in use" >>>> -> >>>> If the new name is already in use >>>> >>>> Is that more correct grammar? Not sure. >>>> >>>> p23: >>>> >>>> "The boot environment has a copy of rpool/zones/z1. The copy is >>>> rpool/zones/z1/ROOT/zbe-1." >>>> >>>> has -> creates >>>> zbe-1 -> zbe >>>> >>>> zbe-1 would be the name of the second zone. >>>> >>>> >>>> "The copy of zone z1 for the new boot environment has its root dataset at >>>> rpool/zones/z1/ROOT/zbe-2" >>>> >>>> zbe-2 -> zbe-1 >>>> >>>> Maybe it would be better to start this example with the original BE being >>>> opensolaris instead of opensolaris-1 and opensolaris-1 being the second >>>> BE, then the zones dataset names won't be conflicting. >>>> >>>> Thanks >>>> Tim >>>> >>>> Barbara.Lundquist at Sun.COM wrote: >>>> >>>> >>>>> Ethan, Evan, Tim, >>>>> >>>>> Here's a final review draft of the snap upgrade document: >>>>> >>>>> http://www.opensolaris.org/os/project/caiman/files/snapupgrade2.pdf >>>>> >>>>> We got alot of useful edits - I think the overall layout of the document >>>>> is >>>>> much improved with the new intro and the overall document more divided >>>>> into "chunks". >>>>> >>>>> Remember this doc will actually be delivered as an HTML document with >>>>> separate HTML pages for each section. Each "how to" section will be on >>>>> its own subpage. >>>>> >>>>> I'll need any final comments by COB this Friday. >>>>> >>>>> Thank you! >>>>> Barbara >>>>> >>>>> _______________________________________________ >>>>> caiman-discuss mailing list >>>>> caiman-discuss at opensolaris.org >>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >>>>> >>>>> >>>> _______________________________________________ >>>> caiman-discuss mailing list >>>> caiman-discuss at opensolaris.org >>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >>>> >>>> >> _______________________________________________ >> caiman-discuss mailing list >> caiman-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >>
