public inbox for pve-devel@lists.proxmox.com
 help / color / mirror / Atom feed
From: Fabian Ebner <f.ebner@proxmox.com>
To: "Fabian Grünbichler" <f.gruenbichler@proxmox.com>,
	"Proxmox VE development discussion" <pve-devel@lists.proxmox.com>
Subject: Re: [pve-devel] [PATCH manager 1/3] close #3476: allow user to prepare storage for activation in job-start hook
Date: Fri, 14 Jan 2022 13:39:10 +0100	[thread overview]
Message-ID: <089ed420-6932-1c47-5281-272cb244a4cf@proxmox.com> (raw)
In-Reply-To: <1642159551.i8ghw8jx3f.astroid@nora.none>

Am 14.01.22 um 12:46 schrieb Fabian Grünbichler:
> while the approach in this series works, I am a bit unhappy about the
> fact that it requires a cfs_update in the (sort of) middle of a worker,
> after having already parsed vmlist and storage.cfg (and making decisions
> based on their contents / saving intermediate results for using later on
> in the process).
> 
> as an alternative, wouldn't calling the hook script either directly
> after forking the vzdump worker (in the API) or somewhere at the start
> of PVE::VZDump->new() work as well? e.g., with a new 'phase' 'job-setup'
> or 'job-init' and just storage/dumpdir and script set. IMHO a sensible
> point would be before the `if ($opts->{storage}) {` in new() that this
> series touches - we have the basic, simple opts handled at that point,
> but haven't started any of the logic/active parts of the initialization
> of the vzdump job, so we are still okay to do a cfs_update without
> worrying about subtle side-effects. and since it would be a new phase,
> we can just document that the storage might not be activated yet at that
> point. obviously this approach is also a bit hacky (run_hookscript
> supposedly takes a PVE::VZDump instance which doesn't exist yet at this
> point ;)), but it seems less involved / with less potential for
> side-effects than the current one?
> 

I wanted to avoid a too similar hook point when there's already 
job-start, but you now convinced me that there's enough need for a 
distinction :)

About the hacky bit: before `if ($opts->{storage}) {`, the instance is 
already blessed, so that shouldn't be an issue either?

> On December 2, 2021 12:00 pm, Fabian Ebner wrote:
>> Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
>> ---
>>   PVE/VZDump.pm | 11 +++++++----
>>   1 file changed, 7 insertions(+), 4 deletions(-)
>>
>> diff --git a/PVE/VZDump.pm b/PVE/VZDump.pm
>> index b5a5fadd..667c92d3 100644
>> --- a/PVE/VZDump.pm
>> +++ b/PVE/VZDump.pm
>> @@ -501,11 +501,8 @@ sub new {
>>   
>>       if ($opts->{storage}) {
>>   	my $storage_cfg = PVE::Storage::config();
>> +	# Ignore errors here. exec_backup will die if activation fails there.
>>   	eval { PVE::Storage::activate_storage($storage_cfg, $opts->{storage}) };
>> -	if (my $err = $@) {
>> -	    chomp($err);
>> -	    $errors .= "could not activate storage '$opts->{storage}': $err";
>> -	}
>>   
>>   	my $info = eval { storage_info ($opts->{storage}) };
>>   	if (my $err = $@) {
>> @@ -1168,6 +1165,12 @@ sub exec_backup {
>>   
>>   	$self->run_hook_script ('job-start', undef, $job_start_fd);
>>   
>> +	my $storage_cfg = PVE::Storage::config();
>> +
>> +	# activating after job-start hook, so it has a chance to prepare, e.g. wake up remote node.
>> +	eval { PVE::Storage::activate_storage($storage_cfg, $opts->{storage}) };
>> +	die "could not activate storage '$opts->{storage}': $@" if $@;
>> +
>>   	foreach my $task (@$tasklist) {
>>   	    $self->exec_backup_task ($task);
>>   	    $errcount += 1 if $task->{state} ne 'ok';
>> -- 
>> 2.30.2
>>
>>
>>
>> _______________________________________________
>> pve-devel mailing list
>> pve-devel@lists.proxmox.com
>> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
>>
>>
>>




  reply	other threads:[~2022-01-14 12:39 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-02 11:00 Fabian Ebner
2021-12-02 11:00 ` [pve-devel] [RFC manager 2/3] vzdump: exec_backup: pick up latest version of storage.cfg after " Fabian Ebner
2021-12-02 11:00 ` [pve-devel] [RFC manager 3/3] vzdump: new: add reminder to get rid of duplicate activate_storage Fabian Ebner
2022-01-14 11:46 ` [pve-devel] [PATCH manager 1/3] close #3476: allow user to prepare storage for activation in job-start hook Fabian Grünbichler
2022-01-14 12:39   ` Fabian Ebner [this message]
2022-01-14 13:07     ` Fabian Grünbichler

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=089ed420-6932-1c47-5281-272cb244a4cf@proxmox.com \
    --to=f.ebner@proxmox.com \
    --cc=f.gruenbichler@proxmox.com \
    --cc=pve-devel@lists.proxmox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal