Koozali.org: home of the SME Server

My experience with SME10 A4

Offline mrjhb3

  • *
  • 1,188
  • +0/-0
    • John Bennett Services
My experience with SME10 A4
« on: October 02, 2019, 12:12:02 PM »

Now that CentOS8 has been released, I gave SME 10a4 a try out.  Installing with single disk worked fine, I was able to create users, run webmail, create ibays.  I didn't try to join a PC to it, as I don't have a spare one to use.  I also couldn't map a drive to it from my W10 workstation, but I swear I had done it before, but maybe not.
 
I was also able to use 2 disks and create a raid1 system.  Swapping the disks, the system booted fine.  If I removed the second disk, the system would boot fine.  If I added files, then booted with both disks in, the second disk would rebuild and be added back to the raid1.  If I removed the first disk, then it wouldn't boot all the way.  It would enter emergency mode.  From what I found, it seemed like it was stopped on some dracut emergency screen.  Trying things from various sites didn't fix the issue for me with booting off the second drive only.I think it was something to do with not knowing the UUID of the second drive, but I won't swear to that.


I was also able to take 3 disks and create a raid5.  /boot was raid1 on the three disks, / and swap were raid 5 using LVM.  If I remove the 1st disk, and put the old 2nd and 3rd disks, 1st and second, then my server boots to the emergency screen with the same UUID missing errors.  Same if I put them back in the original order and remove the 2nd or 3rd disk.


The raid devices were always /md126 and md127.  I found a way to renamed them to /md0 and /md1, but then I wasn't able to reboot after changing them.  There was another dracut command I found on some site, but running those commands didn't fix anything for me, so quit trying to rename them.


Another thing I ran into was an error that I had multiple volumes with the same name.  To fix that, I used the instructions to go to the shell, and use vgrename <UUID of drive> <any_name>.  That would let me proceed.  You can also do a new install on the single disk and change the volume name there.  After the install, you can then use it with other drives.


There were my experiences, and nothing real scientific to what I tried, but hopefully this, and other info provided, will help someone get us further to being able to release SME10.  I don't support SME, I just run it at my home.  I do what I can with the horde webmail product, and all of the changes I have made in 9.2, all seem to work in SME10.  I did add 2 patches for horde in the bug system, so bring things on par with what I have done for 9.2 as well.


Thank you,


John
......

Offline TerryF

  • grumpy old man
  • *
  • 1,821
  • +6/-0
Re: My experience with SME10 A4
« Reply #1 on: October 02, 2019, 12:31:37 PM »
Thanks John for putting in the time and effort to add to the pool..
--
qui scribit bis legit

Offline ReetP

  • *
  • 3,722
  • +5/-0
Re: My experience with SME10 A4
« Reply #2 on: October 02, 2019, 11:12:57 PM »
John,

If you want to come and chat SME (or just life in general!!) then let me know and I can set you up a login on Rocket.Chat where a few of us hang out & chew the cud.

I think we'll be trying to push the development over the forthcoming months and every bit of help is appreciated, even if it just testing.

Same applies to anyone else. Just ask.

...
1. Read the Manual
2. Read the Wiki
3. Don't ask for support on Unsupported versions of software
4. I have a job, wife, and kids and do this in my spare time. If you want something fixed, please help.

Bugs are easier than you think: http://wiki.contribs.org/Bugzilla_Help

If you love SME and don't want to lose it, join in: http://wiki.contribs.org/Koozali_Foundation

Offline bunkobugsy

  • *
  • 274
  • +4/-0
Re: My experience with SME10 A4
« Reply #3 on: October 15, 2019, 10:06:39 AM »
I was also able to use 2 disks and create a raid1 system.  Swapping the disks, the system booted fine.  If I removed the second disk, the system would boot fine.  If I added files, then booted with both disks in, the second disk would rebuild and be added back to the raid1.  If I removed the first disk, then it wouldn't boot all the way.  It would enter emergency mode.  From what I found, it seemed like it was stopped on some dracut emergency screen.  Trying things from various sites didn't fix the issue for me with booting off the second drive only.

The raid devices were always /md126 and md127.  I found a way to renamed them to /md0 and /md1, but then I wasn't able to reboot after changing them. 

Same experience quite a while ago, for me this RAID 1 problem is a showstopper for now (running on simple barebone).

However, I recently installed the just released NethServer 7.7 in unattended mode and it created a functioning RAID1 array (partitions almost identical) that didn't show these symptomps: it booted ok (but slower) with any of the 2 disks missing. Maybe we shold take a look at the unattended kickstart script, managing RAID1 used to work so well in 7,8,9...

Offline ReetP

  • *
  • 3,722
  • +5/-0
Re: My experience with SME10 A4
« Reply #4 on: October 15, 2019, 10:41:27 AM »
Same experience quite a while ago, for me this RAID 1 problem is a showstopper for now (running on simple barebone).

However, I recently installed the just released NethServer 7.7 in unattended mode and it created a functioning RAID1 array (partitions almost identical) that didn't show these symptomps: it booted ok (but slower) with any of the 2 disks missing.

So you had a 3 drive RAID 1?

Cos normally RAID 1 is just a pair of drives.

So you probably meant RAID 5 which is not the same, and will always be slower if you pull a drive as it has to calculate the missing bit via parity.

Don't try comparing apples & oranges. They are different things.

Quote
Maybe we shold take a look at the unattended kickstart script, managing RAID1 used to work so well in 7,8,9...

We? Who do you mean by 'we'?

By all means come and chat to the few of us who do anything and let us know what you want to code.

Be delighted if you can fix some of this.......

We tend to hang about on my Rocket test rig. Ask for a login.

Trying to manage RAID systems automagically has always been fraught.

The issue which SME has suffered has been with grub installation.

That may be easier to manage with CentOS 7 and newer versions of grub.
...
1. Read the Manual
2. Read the Wiki
3. Don't ask for support on Unsupported versions of software
4. I have a job, wife, and kids and do this in my spare time. If you want something fixed, please help.

Bugs are easier than you think: http://wiki.contribs.org/Bugzilla_Help

If you love SME and don't want to lose it, join in: http://wiki.contribs.org/Koozali_Foundation

Offline bunkobugsy

  • *
  • 274
  • +4/-0
Re: My experience with SME10 A4
« Reply #5 on: October 15, 2019, 11:13:43 AM »
Sorry if I've offended any of you.

No, I tested it with 2 identical disks wich I use in Raid1. Periodically I remove disk 2 and replace with a 3rd and rebuild to it (so I always kind of have a booting/working full 'backup'). I would never use Raid 5 as a matter of fact.

Sorry if I opened this discussion in the wrong place. Unfortunately I can't help with coding, but could help with testing.

Offline ReetP

  • *
  • 3,722
  • +5/-0
Re: My experience with SME10 A4
« Reply #6 on: October 15, 2019, 12:12:37 PM »
Sorry if I've offended any of you.

Only when you use the Royal 'We' which always means 'You'.....

Quote
No, I tested it with 2 identical disks wich I use in Raid1. Periodically I remove disk 2 and replace with a 3rd and rebuild to it (so I always kind of have a booting/working full 'backup'). I would never use Raid 5 as a matter of fact.

OK - it wasn't clear from what you said. It will only be slow whilst rebuilding (unsurprisingly). Once rebuilt it should be normal speed. If it isn't then you have another problem.

Quote
Sorry if I opened this discussion in the wrong place. Unfortunately I can't help with coding, but could help with testing.

It is the right place, but you just need to be logical about facts :-)

If you want to chat with us about what to test, how to test etc etc then please ask and I will give you a login to Rocket. It is much easier to casually talk about these things, especially as if we are about we can give you instant feedback to your questions.
...
1. Read the Manual
2. Read the Wiki
3. Don't ask for support on Unsupported versions of software
4. I have a job, wife, and kids and do this in my spare time. If you want something fixed, please help.

Bugs are easier than you think: http://wiki.contribs.org/Bugzilla_Help

If you love SME and don't want to lose it, join in: http://wiki.contribs.org/Koozali_Foundation

Offline bunkobugsy

  • *
  • 274
  • +4/-0
Re: My experience with SME10 A4
« Reply #7 on: October 18, 2019, 12:19:21 PM »
Good news, I've done some more testing with A4:

I followed this guide https://www.clearos.com/resources/documentation/clearos/content:en_us:7_d_raid_1_-_mirroring
created RAID 1 XFS partitions (with 2 disks each) for root (500), swap (4096) and boot the rest (so no LVM and growing?)

At first after applying recent 7.7 updates system wouldn't even boot with latest kernel (3.10.0-1062.1.2.el7.x86_64),
but after applying every update from smeupdates-testing and smetest and waiting for raid to rebuild it started working.

It even successfully boots now (slowly) with any of the 2 disks missing, manually readding partitons rebuilds raid fast.

Joining Win7 to PDC , logging in with admin and browsing share works fine, RSAT can't connect (not even needed?).

Offline Jean-Philippe Pialasse

  • *
  • 2,747
  • +11/-0
  • aka Unnilennium
    • http://smeserver.pialasse.com
Re: My experience with SME10 A4
« Reply #8 on: October 18, 2019, 12:52:48 PM »

Sorry if I opened this discussion in the wrong place. Unfortunately I can't help with coding, but could help with testing.

I am sure you underestimate yourself.  You are able to tes and find workaround and test them.

First part before writing some code.

The second part would be analyzing existent code to find where a modification is needed. That is just reading.  If this is new code there is nothing to do there.

Third part is to document what to do.
Code: [Select]
#create a raid
# if one disk
## do nothing
# else if 2 disks
## create raid 1
#...
Only after comes the « difficult » moment to translate the logic to a strange language called perl or bash.

And like a Molière character, you are already using one of them every day without noticing: bash. So yes you know how to code without knowing it, just  need to add some more words or structures of expression (if/else, while...) by practising your language.