Update without change log...?

[SilverJS]

Not even sure this is a bug, tbh - but it's the first time I've seen this.

I have two freenas boxes, server and backup. I always wait a few days before applying updates, monitoring forums and such, and then only apply the update to my backups box first, testing things out and such.

So. As it stands, my server is on Stable-9.3-201504152200, and I haven't updated it since. Yesterday (May 9th), I decided to upgrade my backups box to Stable-9.3-201505040117, and was going to leave it alone for a day or two before upgrading the server itself. To my surprise, I woke up to an email this morning (following day, May 10th) saying that an update was available for my backups box - but the change log was totally empty!

Checked on GUI - the update tab does show an update is available, with a different alphanumeric tag than my current version, but still no change log. On my server box, if I check for updates, I also get the new alphanumeric tag as the latest version available on the backups box, with the same change log as before.

I'm kind of reluctant to update without knowing what's in the pipe....? Or am I misreading this?

Cheers!

 

[BigDave]

No change log... Is this any different than the doubled change log lists of recent past?
I think I'll wait on applying this update until it (the change log), is explained or corrected.

 

[Ericloewe]

My last changelog worked normally.

 

[dlavigne]

Unfortunately, that seems to be the new way of doing things: https://bugs.freenas.org/projects/f...ions/6cc96125205af310ca52b0a48ba5ad357853e8d6.

 

[jgreco]

Unfortunately, that seems to be the new way of doing things: https://bugs.freenas.org/projects/f...ions/6cc96125205af310ca52b0a48ba5ad357853e8d6.

Yeah? 403 "You are not authorized to access this page" - and yes, I'm logged in.

 

[Mlovelace]

Yeah? 403 "You are not authorized to access this page" - and yes, I'm logged in.

Here is what it says:
« Previous | Next »

Revision 6cc96125
ID6cc96125205af310ca52b0a48ba5ad357853e8d6
Parentc08e9a1f
Child4c5bf4de
Added by Sean Fagan 7 days ago

Change the default changelog to be shown from the full changelog to
nothing.

Ticket: #9559
Merge-FN93: Yes
Merge-TN93: Yes
(cherry picked from commit 56b01d65367d7cab43f8138b8a57e7b0cef31909)

Related issues

• Bug #9559: Flip logic in changelog presentation

 

[SilverJS]

So this is intended behavior then?

 

[BigDave]

Change the default changelog to be shown from the full changelog to nothing.

Intended or not, it's pretty clear... no more change log.

 

[dlavigne]

Not sure if it's a permanent thing or a temporary measure pending a better solution. I'll have more info later this week when I catchup with the latest changes doc-wise. I've been out of the loop for a few weeks doc'ing the PC-BSD stuff in prep for its next release.

 

[SilverJS]

Many thanks Dru!

Might we kindly ask as to the rationale behind this change?

 

[jgreco]

Unfortunately, that seems to be the new way of doing things: https://bugs.freenas.org/projects/f...ions/6cc96125205af310ca52b0a48ba5ad357853e8d6.

Oh cool, I can see it if I sign out of the bugtracker, but I get denied if logged in. Maybe Jordan finally banned me from submitting bugs.

 

[Mlovelace]

Well without knowing what changes are being implemented in an update it makes the decision to schedule an outage difficult.

 

[mjws00]

Just speculating. But I'm guessing the intent was to stop spamming the changelog with every bit of minutae and bugfix. Instead add the important things manually or automate somehow. Perhaps someone in the know will chime in eventually.

 

[Scareh]

Unfortunately this is excactly what I feared would happen eventually when freenas made the switch from the"big version patches/updates" to what we currently have.
Without any clue on whats being changed its impossible for us usera to check if its worth the hassle of possible new bugs being inserted/solved.
Can't even check the forums anymore to see on which update a certain bug would be.
Hopefully someone clarifies/turns back this soon.

 

[SilverJS]

Dru, anything further on this?

 

[Scareh]

any update on this, or a "no comment"? Still curious to the reasoning behind this

 

[Scareh]

my current freenas version:
Build FreeNAS-9.3-STABLE-201505040117
tried another update:

Upgrade: base-os-9.3-STABLE-4c5bf4d-9e81b33-adf8538 -> base-os-9.3-STABLE-78ebdb9-3bda974-b54df29
Upgrade: freebsd-pkgdb-9.3-STABLE-4c5bf4d-9e81b33-adf8538 -> freebsd-pkgdb-9.3-STABLE-78ebdb9-3bda974-b54df29
Upgrade: FreeNASUI-9.3-STABLE-4c5bf4d-9e81b33-adf8538 -> FreeNASUI-9.3-STABLE-78ebdb9-3bda974-b54df29
Upgrade: freenas-pkg-tools-9.3-STABLE-4c5bf4d -> freenas-pkg-tools-9.3-STABLE-78ebdb9
Upgrade: docs-9.3-STABLE-ccbc6e6 -> docs-9.3-STABLE-78ebdb9

and when i click the changelog button:


http://update.freenas.org/FreeNAS/FreeNAS-9.3-STABLE/Notes/ChangeLog-fpNm35.txt

Code:

#7115 Bug Nice to have VAAI TP STUN fixed for quota overflows
#8955 Bug Nice to have UFS Disk Import Progress not working
#9255 Bug Nice to have Adds support for service-oriented updates, which may not require a reboot.
#9345 Bug Expected Hide username/passwords for CHAP iSCSI in save debug output
#9417 Bug No priority Fix SMART status checking/fixing in ix-smartd
Add alertmod for monitoring smartd's status
Restart smartd after adding or replacing a drive to enable monitoring for it
#9428 Bug Important Adds support for service-oriented updates, which may not require a reboot.
#9546 Bug Nice to have 201505010007 Update Killed my CIFS Share
#9556 Bug Nice to have Fix line break for smart alerts
#9562 Bug Important Force system dataset to to be of type 'unix' (specifical, don't set aclmode to restricted). Use Thor hammer to do so.
#9565 Bug Nice to have CIFS->workgroup cannot accept 1 char length
#9708 Bug Important Increase smartd's read buffer to allow for up to 512 characters so more complex SMART test schedules will work proper
#9715 Bug Expected Fixes where ChangeLog is not be proper appended to for display.

so i'm thinking this has been reversed without any comment.

 

[jkh]

You guys are all wrong

Sorry, I just thought I'd get that out of the way first. :)

The ChangeLog behavior has been pretty much exactly what one would expect it to be for quite some time now (no bugs open against it, and at least the last 5 updates have been "ChangeLog all good, thumbs up!"). Each SU contributes a chunk of changelog text, and depending on how old your OS version is, you'll get all of the relevant chunks concatenated. This was never turned off or changed. There were a few bugs which scrambled (or outright disabled) the ChangeLog contents early in 9.3's release cycle, but those were bugs. They got fixed.

The git change that Dru is referring to has nothing to do with this. It changed the way that ChangeLogs used to be dumped en-masse at you when the user switched trains. Now, switching update trains simply starts the user over from the new ChangeLog train rather than spewing all accumulated ChangeLog data from the dawn of time, which was never the right behavior (arguably also a bug).

HTH