software or hardware raid?

[Ranjan Maitra]

Hi,

I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Thanks,
Ranjan

[Tom Horsley]

On Wed, 1 Mar 2023 14:01:37 -0600
Ranjan Maitra wrote:
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

I'm pretty sure software is better. You don't get weird hardware
implementations which make your disks inaccessible if the hardware
breaks and you can't replace it with identical hardware. Modern
computers are so fast there really isn't a speed penalty for
software raid any longer (not one a human could notice anyway).

[John Mellor]

On 2023-03-01 3:01 p.m., Ranjan Maitra wrote:
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Don't even consider hardware RAID solutions that became out-of-favour 25 
years ago.  You can easily do it with LVM/MDR combinations, but its more 
complicated and slower than the BTRFS solution, especially when 
rebuilding onto a new drive.  Instead, because the normal Fedora 
installer is a bit brain-dead, unplug one of the drives and use the 
Fedora-default BTRFS setup.  Then plug in the second drive back again, 
reboot and set up a RAID-1 configuration.  See 
http://www.beginninglinux.com/btrfs for a quick tutorial.  There is also 
a good reference guide on kernel.org, at 
https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices 
to see what can be done.  BTRFS is massively faster and safer than all 
other implementations other than perhaps ZFS, is at least partially 
error compensating, does not demand identical drives, and can be easily 
converted into other RAID classes dynamically as desired.

[Richard Shaw]

On Wed, Mar 1, 2023 at 2:45 PM John Mellor  wrote:
> On 2023-03-01 3:01 p.m., Ranjan Maitra wrote:
> > I would like to RAID two of my disks, and I was wondering if the
>
> recommendation is to do software or hardware RAID?
> Don't even consider hardware RAID solutions that became out-of-favour 25
> years ago.  You can easily do it with LVM/MDR combinations, but its more
> complicated and slower than the BTRFS solution, especially when
> rebuilding onto a new drive.  Instead, because the normal Fedora
> installer is a bit brain-dead, unplug one of the drives and use the
> Fedora-default BTRFS setup.  Then plug in the second drive back again,
> reboot and set up a RAID-1 configuration.  See
> http://www.beginninglinux.com/btrfs for a quick tutorial.  There is also
> a good reference guide on kernel.org, at
> https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices
> to see what can be done.  BTRFS is massively faster and safer than all
> other implementations other than perhaps ZFS, is at least partially
> error compensating, does not demand identical drives, and can be easily
> converted into other RAID classes dynamically as desired.

I was going to recommend BTRFS but there's still a lot of people weary
about it because of the long time to stabilization.

I have 4 4TB drives in a BTRFS RAID 1 array for all my media. I know RAID 5
is a bit touchy so I have put that off until I start to run out of space.
The cool thing is I can convert on the fly!

I have installed btrfsmaintenancetools or whatever the package name is and
am running scrubs and balances regularly.

Thanks,
Richard

[Slade Watkins]

On 3/1/23 15:01, Ranjan Maitra wrote:
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Honestly, in my experience, I've found software RAIDs to be much more reliable
than hardware. But as always, your mileage may vary.

-- Slade

[Peter Boy]

> Am 01.03.2023 um 21:45 schrieb John Mellor :
>
> BTRFS is massively faster and safer than all other implementations other than perhaps ZFS, is at least partially error compensating, does not demand identical drives, and can be easily converted into other RAID classes dynamically as desired.

That's a strong claim.  And it is basically as false as it is strong.

There is a lot of discussion about BTRFS regarding performance and stability. And because BTRFS is so superb fast and reliable, Red Hat has dropped BTRFS in RHEL. They don’t want something so fast and reliable in their distro. This would probably make the (paid) support redundant and reduce the profit. And customers could become overwhelmed and dissatisfied with so much speed. </irony>

BTRFS has a number of advantages, but also some disadvantages. A slightly more accurate and less operationally blind presentation would be much more appropriate. As is almost always the case with Technology, it's a matter of balance and differing relevancies.

--
Peter Boy
https://fedoraproject.org/wiki/User:Pboy
[e-mail]

Timezone: CET (UTC+1) / CEST (UTC+2)

Fedora Server Edition Working Group member
Fedora docs team contributor
Java developer and enthusiast

[Roger Heflin]

On Wed, Mar 1, 2023 at 2:02 PM Ranjan Maitra  wrote:
> Hi,
>
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?
>
> Thanks,
> Ranjan

I work with a lot of hardware raid from a couple of vendors.

The documentation is rather sparse and when it does exist not always
consistent and/or complete.

And if something goes wrong with the firmware (like the array drops
offline all at once) the recovery is not always well documented.

And that ignores the fact that if your hw controller dies that you
need to find a compatible hardware controller that can make sense of
the raid format saved on the disks.

My current array is 10 years old and has been moved to 3 different
motherboards and has used at least 3 different add-on SATA controller
cards.

[John Mellor]

On 2023-03-01 4:37 p.m., Peter Boy wrote:
> > Am 01.03.2023 um 21:45 schrieb John Mellor :
> >
> > BTRFS is massively faster and safer than all other implementations other than perhaps ZFS, is at least partially error compensating, does not demand identical drives, and can be easily converted into other RAID classes dynamically as desired.
>
> That's a strong claim.  And it is basically as false as it is strong.
>
> ...

In what way is it incorrect?  You should try maybe benchmarking common 
operations as well as some actual rebuilds before making this incorrect 
claim about being false.  I'd recommend ZFS as being even better, but a 
decent implementation like on Ubuntu is not available in Fedora.

[Roberto Ragusa]

On 3/1/23 21:01, Ranjan Maitra wrote:
> Hi,
>
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Software.
Everything you need is in the mdadm command.

Regards.

-----Original Message-----
From: Ranjan Maitra 
Sent: Wednesday, March 1, 2023 9:02 PM
To: Community Support for Fedora Users 
Subject: software or hardware raid?

Hi,

I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Thanks,
Ranjan
_______________________________________________
users mailing list -- [e-mail] To unsubscribe send an email to [e-mail]
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

In most cases, i would say: soft raid
- nowadays CPU's can do easily the workload, compared to 20 years ago,
- no dependency on specific hardware

If you had a g HP-server, with a vast amount of drives,  and wanting raid-60, and having enough spare-parts, you could consider HW-raid.

Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u niet de geadresseerde bent of dit bericht abusievelijk aan u is toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht te verwijderen. De Staat aanvaardt geen aansprakelijkheid voor schade, van welke aard ook, die verband houdt met risico's verbonden aan het elektronisch verzenden van berichten.

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. The State accepts no liability for damage of any kind resulting from the risks inherent in the electronic transmission of messages.

On 2023-03-01 12:01 pm, Ranjan Maitra wrote:
> Hi,
>
> I would like to RAID two of my disks, and I was wondering if the
> recommendation is to do software or hardware RAID?
>
> Thanks,
> Ranjan
> _______________________________________________
> users mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct:
> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives:
> https://lists.fedoraproject.org/archives/list/[e-mail]
> Do not reply to spam, report it:
> https://pagure.io/fedora-infrastructure/new_issue

Hi;

I've a btrfs RAID1 array comprised of two 2tb hard drives up and running
for just under a year.  It's mounted as /home and backed up regularly
to a external usb hard drive.  I've thought of upgrading to ssd drives
in the PCI slots though.

Joe

[George N. White III]

On Wed, Mar 1, 2023 at 4:02 PM Ranjan Maitra  wrote:
> Hi,
>
> I would like to RAID two of my disks, and I was wondering if the
> recommendation is to do software or hardware RAID?

Software RAID works very well on modern hardware.  You do want to make sure
a power
outage can't cause a shutdown before the RAID is safely dismounted.
Ideally a your site
has a reliable generator and your server has a UPS that can hold power
while waiting for
the generator to come online and also long enough to ensure a clean
shutdown when the
generator fails.

--
George N. White III

[Ranjan Maitra]

Thanks to everybody. I recall discussion from several years ago on the benefits of software over hardware RAID. I had completely forgotten about UPS for this new machine. Btw, what happens if power goes out (and I do not have UPS)?

Btw, I still stick to ext4, largely because of inertia (and because I have used lvm in the past and hated its naming conventions, I think, but there were also other limitations that I do not now recall) and have stayed away from zfs or btrfs or lvm. I am not sure what to do now. Clearly, things have moved far on.

Thanks,
Ranjan

On Thu Mar02'23 02:19:25PM, George N. White III wrote:
> From: "George N. White III" 
> Date: Thu, 2 Mar 2023 14:19:25 -0400
> To: Community support for Fedora users 
> Reply-To: Community support for Fedora users 
> Subject: Re: software or hardware raid?
>
> On Wed, Mar 1, 2023 at 4:02 PM Ranjan Maitra  wrote:
> > Hi,
> >
> > I would like to RAID two of my disks, and I was wondering if the
> > recommendation is to do software or hardware RAID?
>
> Software RAID works very well on modern hardware.  You do want to make sure
> a power
> outage can't cause a shutdown before the RAID is safely dismounted.
> Ideally a your site
> has a reliable generator and your server has a UPS that can hold power
> while waiting for
> the generator to come online and also long enough to ensure a clean
> shutdown when the
> generator fails.
>
> --
> George N. White III
> users mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]
> Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

[Richard Shaw]

On Thu, Mar 2, 2023 at 2:25 PM Ranjan Maitra  wrote:
> Thanks to everybody. I recall discussion from several years ago on the
> benefits of software over hardware RAID. I had completely forgotten about
> UPS for this new machine. Btw, what happens if power goes out (and I do not
> have UPS)?

If you don't have a UPS, I would recommend BTRFS mostly because it's a
Copy-On-Write (COW) filesystem.  EXT4 journaling only protects the
filesystem, not the files themselves. BRTFS does a checksum of the files as
well, and because it writes out a new file, if that's interrupted, you
still at least have the previous version of the file.

Depending on the type of files, you may also want to take advantage of
transparent compression. I know BTRFS took a long time to stabilize but it
works pretty well for stand alone and RAID 1. I'm not sure if I would trust
it for RAID 5 just yet.

Thanks,
Richard

[Roger Heflin]

if you suddenly lose power there is a fair chance that the last few
blocks of data had not made it to disk yet.

For the most part this only only a big issue with oracle db and/or
mysql and/or stuff with critical transactions that cannot be lost and
that need a consistent state when they come back up and flush writes
and/or use direct io.

For anything else the bit of data you lost is not so much an issue
since whatever that was stopped working and did not finish.  I run
cameras and other stuff, and the 5-30 seconds I could lose because of
this is minor compared to what I lost for being down for however long
I was down for.

On Thu, Mar 2, 2023 at 2:25 PM Ranjan Maitra  wrote:
> Thanks to everybody. I recall discussion from several years ago on the benefits of software over hardware RAID. I had completely forgotten about UPS for this new machine. Btw, what happens if power goes out (and I do not have UPS)?
>
> Btw, I still stick to ext4, largely because of inertia (and because I have used lvm in the past and hated its naming conventions, I think, but there were also other limitations that I do not now recall) and have stayed away from zfs or btrfs or lvm. I am not sure what to do now. Clearly, things have moved far on.
>
> Thanks,
> Ranjan
>
> On Thu Mar02'23 02:19:25PM, George N. White III wrote:
> > From: "George N. White III" 
> > Date: Thu, 2 Mar 2023 14:19:25 -0400
> > To: Community support for Fedora users 
> > Reply-To: Community support for Fedora users 
> > Subject: Re: software or hardware raid?
> >
> > On Wed, Mar 1, 2023 at 4:02 PM Ranjan Maitra  wrote:
> > > Hi,
> > >
> > > I would like to RAID two of my disks, and I was wondering if the
> > > recommendation is to do software or hardware RAID?
> >
> > Software RAID works very well on modern hardware.  You do want to make sure
> > a power
> > outage can't cause a shutdown before the RAID is safely dismounted.
> > Ideally a your site
> > has a reliable generator and your server has a UPS that can hold power
> > while waiting for
> > the generator to come online and also long enough to ensure a clean
> > shutdown when the
> > generator fails.
> >
> > --
> > George N. White III
> > users mailing list -- [e-mail]
> > To unsubscribe send an email to [e-mail]
> > Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> > List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]
> > Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
>
> users mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]
> Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

[Chris Adams]

Once upon a time, Ranjan Maitra  said:
> Thanks to everybody. I recall discussion from several years ago on the benefits of software over hardware RAID. I had completely forgotten about UPS for this new machine. Btw, what happens if power goes out (and I do not have UPS)?

Linux software RAID keeps a bitmap of pending writes by default, which
is an okay (but not perfect) mechanism to recover from unexpected
shutdown.  There's also an option to keep a write journal instead, but
unless you put that on a separate fast device (e.g. quality SSD with
long write lifetime), it'll impact performance significantly.

There are trade-offs between various types of SW and HW RAID, so really
the first question would be "what are your requirements and
expectations".  Are you talking about a high-uptime server, or a desktop
where you just want to make hardware failure less annoying?  RAID (HW or
SW) is NOT backups, so you shouldn't depend on it for saving your data.

HW RAID has some advantages - quality controllers will have
battery-backed cache, so things like write journaling don't impact
performance and recovery from unexpected power failures is basically
instantaneous.  For high performance requirements, there's less overhead
with HW RAID (because data only has to transit the bus once, then the
RAID controller has its own paths to the drives).  But HW RAID typically
requires odd and/or proprietary software to manage, detect failures,
etc.  Depending on the RAID level you are using, recovery from a failure
of the controller itself can be harder too.
> Btw, I still stick to ext4, largely because of inertia (and because I have used lvm in the past and hated its naming conventions, I think, but there were also other limitations that I do not now recall) and have stayed away from zfs or btrfs or lvm. I am not sure what to do now. Clearly, things have moved far on.

I'm generally in the XFS on LVM (on SW mdraid when needed) camp
myself... LVM adds a significant layer of flexibility and ability, but
still using more "traditional" filesystems like XFS and ext4.  I had
poor experiences with ZFS at a former job, and am still a little leery
of some of the approach BTRFS takes.

I'm playing with adding the dm-integrity layer for my SW mdraid (so then
XFS on LVM on mdraid on integrity on drive) setup as an additional check
against silent drive failures, but again, unless you put that data on a
separate fast SSD, it slows down performance a lot.

[Ranjan Maitra]

Thanks, Chris!

On Thu Mar02'23 02:49:49PM, Chris Adams wrote:
> From: Chris Adams 
> Date: Thu, 2 Mar 2023 14:49:49 -0600
> To: [e-mail]
> Reply-To: Community support for Fedora users 
> Subject: Re: software or hardware raid?
>
> Once upon a time, Ranjan Maitra  said:
> > Thanks to everybody. I recall discussion from several years ago on the benefits of software over hardware RAID. I had completely forgotten about UPS for this new machine. Btw, what happens if power goes out (and I do not have UPS)?
>
> Linux software RAID keeps a bitmap of pending writes by default, which
> is an okay (but not perfect) mechanism to recover from unexpected
> shutdown.  There's also an option to keep a write journal instead, but
> unless you put that on a separate fast device (e.g. quality SSD with
> long write lifetime), it'll impact performance significantly.
>
> There are trade-offs between various types of SW and HW RAID, so really
> the first question would be "what are your requirements and
> expectations".  Are you talking about a high-uptime server, or a desktop
> where you just want to make hardware failure less annoying?  RAID (HW or
> SW) is NOT backups, so you shouldn't depend on it for saving your
> data.

Thanks, this will be a fairly high uptime machine (not allowed to call it a server here, because that is central IT's role to have and administer:-), running lots of jobs at least a large part of the time, but the  RAID will be on the /. It is more to keep the machine going if one of the two / drives fail (and till such time as I can get and put in a new one).

For /home (which is where my data reside), I have 2 backups done using rsync every hour. I plan to copy the actual /home to the second one, and I was  thinking that the third one would be incremental backup (kept  for a year, since I occasionally realize weeks and months later that I really want a file back from long ago) or so.
> HW RAID has some advantages - quality controllers will have
> battery-backed cache, so things like write journaling don't impact
> performance and recovery from unexpected power failures is basically
> instantaneous.  For high performance requirements, there's less overhead
> with HW RAID (because data only has to transit the bus once, then the
> RAID controller has its own paths to the drives).  But HW RAID typically
> requires odd and/or proprietary software to manage, detect failures,
> etc.  Depending on the RAID level you are using, recovery from a failure
> of the controller itself can be harder too.
> > Btw, I still stick to ext4, largely because of inertia (and because I have used lvm in the past and hated its naming conventions, I think, but there were also other limitations that I do not now recall) and have stayed away from zfs or btrfs or lvm. I am not sure what to do now. Clearly, things have moved far on.
>
> I'm generally in the XFS on LVM (on SW mdraid when needed) camp
> myself... LVM adds a significant layer of flexibility and ability, but
> still using more "traditional" filesystems like XFS and ext4.  I had
> poor experiences with ZFS at a former job, and am still a little leery
> of some of the approach BTRFS takes.
>
> I'm playing with adding the dm-integrity layer for my SW mdraid (so then
> XFS on LVM on mdraid on integrity on drive) setup as an additional check
> against silent drive failures, but again, unless you put that data on a
> separate fast SSD, it slows down performance a lot.

I see, so your recommendation is to go for xfs?

Many thanks again, and best wishes,
Ranjan

[Chris Adams]

Once upon a time, Ranjan Maitra  said:
> Thanks, this will be a fairly high uptime machine (not allowed to call it a server here, because that is central IT's role to have and administer:-), running lots of jobs at least a large part of the time, but the  RAID will be on the /. It is more to keep the machine going if one of the two / drives fail (and till such time as I can get and put in a new one).

That's a good target for RAID (I just like to remind people RAID is not
backups, because double drive failures happen, filesystem corruption
happens, somebody deletes the wrong file happens, and so on).
> I see, so your recommendation is to go for xfs?

It's what I use, in part because I also run RHEL and CentOS servers,
where XFS is the default and preferred filesystem by Red Hat for a while
now.

I think the pros and cons of XFS vs ext4 probably aren't that
significant in most use.  XFS doesn't currently support any kind of
shrink operation (more of an issue if you are using LVM but not LVM thin
pools, and there is some work on adding this).  ext4 can also journal
data (doesn't by default by can be enabled), which gives additional
protection (at an additional performance cost).  XFS is higher
performance for some uses, but that probably gets into specifics about
your use cases to know if it really is (or if it matters).  XFS supports
reflinks while ext4 does not, which again can be useful for certain
things.

If you are familiar and happy with ext4 though, there's no reason to
switch unless you see something in particular that XFS would do better
in your use.  ext4 is not going away any time soon, and both ext4 and
XFS are mature and stable filesystems (and both are still getting
development).

[George N. White III]

On Thu, Mar 2, 2023 at 4:43 PM Richard Shaw  wrote:
> On Thu, Mar 2, 2023 at 2:25 PM Ranjan Maitra  wrote:
> > Thanks to everybody. I recall discussion from several years ago on the
> > benefits of software over hardware RAID. I had completely forgotten about
> > UPS for this new machine. Btw, what happens if power goes out (and I do not
> > have UPS)?
>
> If you don't have a UPS, I would recommend BTRFS mostly because it's a
> Copy-On-Write (COW) filesystem.  EXT4 journaling only protects the
> filesystem, not the files themselves. BRTFS does a checksum of the files as
> well, and because it writes out a new file, if that's interrupted, you
> still at least have the previous version of the file.
>
> Depending on the type of files, you may also want to take advantage of
> transparent compression. I know BTRFS took a long time to stabilize but it
> works pretty well for stand alone and RAID 1. I'm not sure if I would trust
> it for RAID 5 just yet.

Here in Canada, we often get short power hits, often from a vehicle sliding
on snow or ice into a power pole. That
often causes widespread loss-of-power just long enough to stop computers.
UPS means jobs keep going instead
waiting for reboot and filesystem repairs, followed by restarting jobs that
otherwise would be finished.  Sometimes
systems don't come up without added work.

[George N. White III]

On Thu, Mar 2, 2023 at 4:49 PM Roger Heflin  wrote:
> if you suddenly lose power there is a fair chance that the last few
> blocks of data had not made it to disk yet.

Some data can be replaced, but there are also real-time data flows
where data are lost forever when the system does down.  I recall a
lost-of-data incident where the system was on UPS+ generator but
IT hadn't received the replacement for a failed UPS battery in the
network closet.

[Chris Adams]

Once upon a time, George N. White III  said:
> Some data can be replaced, but there are also real-time data flows
> where data are lost forever when the system does down.  I recall a
> lost-of-data incident where the system was on UPS+ generator but
> IT hadn't received the replacement for a failed UPS battery in the
> network closet.

Lots of the time, UPSes and generators are not actively monitored and
tested.  An untested backup system is not a backup system, it's just
another point of failure!  I have seen failures of big UPSes,
generators, transfer switches... you name it, even if it is "redundant",
it can (and will) still fail.

[Jeffrey Walton]

On Wed, Mar 1, 2023 at 3:02 PM Ranjan Maitra  wrote:
> I would like to RAID two of my disks, and I was wondering if the recommendation is to do software or hardware RAID?

Also see "A Journal for MD/RAID5", https://lwn.net/Articles/665299/
and "ext3 and RAID: silent data killers?",
https://lwn.net/Articles/349970/ . But both articles are kind of old.

Jeff

[Tim]

On Thu, 2023-03-02 at 17:09 -0600, Chris Adams wrote:
> Lots of the time, UPSes and generators are not actively monitored and
> tested.  An untested backup system is not a backup system, it's just
> another point of failure!  I have seen failures of big UPSes,
> generators, transfer switches... you name it, even if it is "redundant",
> it can (and will) still fail.

I have a UPS sitting next to me, right now, in pieces, which (half)
died in a most peculiar manner:

A burning smell was eventually traced to it.  There's no visible signs
of burning, and no schematic available for the model, that I can find. 
A rather acrid smell, not one I'm used to with component failure, I'm
beginning to suspect a large AC transformer.

While running off the mains its output is a (too) low voltage, but
still high enough for most switch-mode power supplies to run normally
(i.e. the computer and monitor).  It has some kind of AC voltage
regulation built into it to deal with under and over-voltage.  The
AC supply was normal, at the time.

But running off its battery it produces the full 240 volts it's
supposed to.

[Mauricio Tavares]

On Thu, Mar 2, 2023 at 10:42 PM Tim via users
 wrote:
> On Thu, 2023-03-02 at 17:09 -0600, Chris Adams wrote:
> > Lots of the time, UPSes and generators are not actively monitored and
> > tested.  An untested backup system is not a backup system, it's just
> > another point of failure!  I have seen failures of big UPSes,
> > generators, transfer switches... you name it, even if it is "redundant",
> > it can (and will) still fail.
>
> I have a UPS sitting next to me, right now, in pieces, which (half)
> died in a most peculiar manner:
>
> A burning smell was eventually traced to it.  There's no visible signs
> of burning, and no schematic available for the model, that I can find.
> A rather acrid smell, not one I'm used to with component failure, I'm
> beginning to suspect a large AC transformer.

Did you check the impedance of the transformer?
> While running off the mains its output is a (too) low voltage, but
> still high enough for most switch-mode power supplies to run normally
> (i.e. the computer and monitor).  It has some kind of AC voltage
> regulation built into it to deal with under and over-voltage.  The
> AC supply was normal, at the time.
>
> But running off its battery it produces the full 240 volts it's
> supposed to.
>
> --
>
> uname -rsvp
> Linux 3.10.0-1160.83.1.el7.x86_64 #1 SMP Wed Jan 25 16:41:43 UTC 2023 x86_64
>
> Boilerplate:  All unexpected mail to my mailbox is automatically deleted.
> I will only get to see the messages that are posted to the mailing list.
>
> users mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]
> Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

[Tim]

Tim:
> > I have a UPS sitting next to me, right now, in pieces, which (half)
> > died in a most peculiar manner:
> >
> > A burning smell was eventually traced to it.  There's no visible
> > signs of burning, and no schematic available for the model, that I
> > can find.  A rather acrid smell, not one I'm used to with component
> > failure, I'm beginning to suspect a large AC transformer.

Mauricio Tavares:
> Did you check the impedance of the transformer?

Not yet, the lack of schematics put me off, I'd really have to delve
into reverse-engineering it to fix it (assuming I could get any needed
parts), it's a very old model (SOLA 310).  If the transformer had
cooked one of its windings, I doubt I'd find replacement (there are
multiple windings), and the board is chock full of ICs (very few
discrete components).

I never really bothered with UPSs at home before, but we've been
getting lots of little power cuts the last year.  Previously things
were very good here.  You might have gone a year or more without any
interruptions.

The other thing is that this only has a few minutes of supply
capability, allowing you to ride through blips on the mains, and
cleanly shutdown during longer power failures, but not carry on working
for a prolonged period.  Though my modern low-power PC gives much more
runtime than the UPS's prediction in its manual.

And it was quite noisy in stand-by mode, while AC powered, it had quite
a hum to it.  Bad enough that I'd run it in the adjacent room, with
long leads running to and from it.

Much as I dislike waste, and I repair a lot of equipment, I've been
considering something newer with more capacity.

[Roger Heflin]

On Thu, Mar 2, 2023 at 9:42 PM Tim via users
 wrote:
> I have a UPS sitting next to me, right now, in pieces, which (half)
> died in a most peculiar manner:
>
> A burning smell was eventually traced to it.  There's no visible signs
> of burning, and no schematic available for the model, that I can find.
> A rather acrid smell, not one I'm used to with component failure, I'm
> beginning to suspect a large AC transformer.
>
> While running off the mains its output is a (too) low voltage, but
> still high enough for most switch-mode power supplies to run normally
> (i.e. the computer and monitor).  It has some kind of AC voltage
> regulation built into it to deal with under and over-voltage.  The
> AC supply was normal, at the time.
>
> But running off its battery it produces the full 240 volts it's
> supposed to.

If you read about normal UPSes they are not normally designed to run
100% duty cycle (ie on battery for days, or fixing up a low voltage
for days).  So if you run yours at say 40% load it will probably
survive under the higher duty cycle, but if you run it close to 100%
load and it gets into cleaning up low voltages the UPS electronics may
not survive long (ie if fixing the voltages for hours/days).

So if they end up running in that state for long periods of time
(either because of adding bigger batteries--my UPS came with 12AH
batterys but now has external 35AH ones) and/or extended low voltages
various components may burn out.

There is also a setting on at least some UPSes that you can change so
that it does not regulate the lower voltages (wider acceptable voltage
range) and accepts those voltages as ok. I have done that with both of
my UPSes because the default setting prevents the UPSes from charging
when on a small generator, and would if the voltages was low but still
good enough cause the UPS to keep fixing the slightly low voltages and
possibly burn out.

[Jeffrey Walton]

On Fri, Mar 3, 2023 at 7:54 AM Roger Heflin  wrote:
> [...]
>
> If you read about normal UPSes they are not normally designed to run
> 100% duty cycle (ie on battery for days, or fixing up a low voltage
> for days).  So if you run yours at say 40% load it will probably
> survive under the higher duty cycle, but if you run it close to 100%
> load and it gets into cleaning up low voltages the UPS electronics may
> not survive long (ie if fixing the voltages for hours/days).
>
> So if they end up running in that state for long periods of time
> (either because of adding bigger batteries--my UPS came with 12AH
> batterys but now has external 35AH ones) and/or extended low voltages
> various components may burn out.
>
> There is also a setting on at least some UPSes that you can change so
> that it does not regulate the lower voltages (wider acceptable voltage
> range) and accepts those voltages as ok. I have done that with both of
> my UPSes because the default setting prevents the UPSes from charging
> when on a small generator, and would if the voltages was low but still
> good enough cause the UPS to keep fixing the slightly low voltages and
> possibly burn out.

I don't think the part about a generator is a good idea. Generators
are usually wildly out of spec for Total Harmonic Distortion (THD). A
generator under load can easily reach 15% to 20% THD. Meanwhile,
electronics usually expect 3% THD or less. So you want the signal
conditioned, and not passing directly through to the electronics.

Jeff

[Roger Heflin]

On Fri, Mar 3, 2023 at 7:07 AM Jeffrey Walton  wrote:
> I don't think the part about a generator is a good idea. Generators
> are usually wildly out of spec for Total Harmonic Distortion (THD). A
> generator under load can easily reach 15% to 20% THD. Meanwhile,
> electronics usually expect 3% THD or less. So you want the signal
> conditioned, and not passing directly through to the electronics.
>
> Jeff

I have ran computers on wild generators for several days and no issues.

The switching power supplies are very good at taking just about any
power with a reasonable voltage and just working.   They take the
input voltages and convert to DC pulses and really don't care much
about the input so long as the HZ is above the rated HZ (usually 50
for non-US), and the voltage is not really high or really low.

The only real issue is if the input voltage is significantly low and
the output power is a high percentage of rated and the lower input
voltage causes current draws that are at the high end of design.

The switching power supply really eliminates a lot of the input power
requirements (in computers).

If the electronics do not have a switching power supply then linear
power supplies are a lot less forgiving.  In a linear power supply the
transformer is first, in a switching the transformer is after the DC
pulses so really does not care about wide frequency variations and
voltage variations nearly as much.

[Tim]

Tim:
> > I have a UPS sitting next to me, right now, in pieces, which (half)
> > died in a most peculiar manner:
> > While running off the mains its output is a (too) low voltage, but
> > still high enough for most switch-mode power supplies to run
> > normally (i.e. the computer and monitor).  It has some kind of AC
> > voltage regulation built into it to deal with under and over-
> > voltage.  The AC supply was normal, at the time.
> >
> > But running off its battery it produces the full 240 volts it's
> > supposed to.

Roger Heflin:
> If you read about normal UPSes they are not normally designed to run
> 100% duty cycle (ie on battery for days, or fixing up a low voltage
> for days).  So if you run yours at say 40% load it will probably
> survive under the higher duty cycle, but if you run it close to 100%
> load and it gets into cleaning up low voltages the UPS electronics may
> not survive long (ie if fixing the voltages for hours/days).

Its a Sola 0310-0400 A.  The UPS is specced for 188 to 264 Vac in,
intended for a normally 220-240 volt AC supply.  Outputs a modified
sinewave,  of 230 Vac up to 2.35 amps max, 1.91 amps nominal, 400 watts
(VA) output, or 250 watts (taking power factor into consideration). 
Able to power a full load for 4 minutes, 15 minutes for a half load. 
Which all sounds like tax accountant maths, pitched by a used car
salesman.

I live in a 240 volt country, the mains here is generally 240 to 245
volts.  In general, it just sits there coasting along on easy street. 
The mains is close to spec 99% of the time.  The computer only uses
about 40 watts, the monitor a bit less (when it's on, next to nothing
when off, which is mostly is).  Light duties with a very basic system.

The UPS draws about 16 watts from the wall when running without a
computer attached (maintaining its fully charged battery, I haven't
tested it recharging a flattened one).  And always humming like it went
deaf at Woodstock (it's always been a loud hummer), with extra buzzing
when its running off its battery.

As I recall, when I assessed it years ago it did what I expected it to
do, as far as output voltages are concerned.  After the recent failure,
the AC output when mains power is present is around 180 volts, but
returns to 240 when running on battery.  That was tested with a
computer as a load, and then separately with an 60 watt incandescent
load (the simplest load I could think of to non-stress test it).

While 180 volts is enough to power the computer equipment without them
being stressed about it (they use universal power supplies meant for
110 to 240 volt countries).  It's clearly faulty, and I tend to think
it's done its duty and earnt its retirement.
> There is also a setting on at least some UPSes that you can change

Unfortunately, it's got no user controls (beyond the on/off and
start/test switches), nor any internal adjustments.  The only
changeable thing is an unidentified 3 pin header with a jumper on it.

I could only venture a guess that the jumpered one is possibly a 50/60
Hz default output selector.  Ordinarily it detects what the mains does,
and follows that.  But a cold start running off the battery with no
mains always starts running at 50 Hz.

While I can easily test moving that jumper, I'm not going to do
potentially nasty electrical work at 1 am.

Now for the kicker, having said I won't mess with mains in the wee
hours:  I forgot about the jumper, but I just poked around with the UPS
now, measured some unwanted resistance on the relay changeover
terminals.  Unsoldered it, poked at the terminals by themselves, it got
better.  Resoldered it back in.  Mains power goes through at full
voltage, battery generated AC output is low.  I may just have a crapped
relay.  It's a potted one, where you can't see the contacts.  That's a
part I can easily replace to see what happens.

[Tim]

On Thu, 2023-03-02 at 19:36 -0500, Jeffrey Walton wrote:
> Also see "A Journal for MD/RAID5", https://lwn.net/Articles/665299/
> and "ext3 and RAID: silent data killers?",
> https://lwn.net/Articles/349970/0/ . But both articles are kind of old.

Yes, is a 2009 article still valid?  Likewise with the 2015 one?

I did read the articles, but not the comments.  With the exception of
one, they're as old as the articles.