• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Please let me know if the UPS that I want to buy is good or not for my PC.

I see those errors when I copy files from and there a ZFS disk. I'm experiencing damages on and to the ZFS disks.

How is the drive powered?
 
Two things - first, post the complete dmesg log, it is important to see which error occurred first. Ubuntu keeps the full log in /var/log/dmesg.
Second, I would suggest to remove devices one by one to get to the state when you don't see any errors.
One other thing to check - is the USB drive power supply plugged in, and is it plugged into the UPS as well ? If you electrical circuits have issues that having computer and USB drive powered from different outlets can have problems. Also for protection from surges you would want all devices electrically connected with the computer to plugged into the UPS (your UPS probably has dedicated outlets with no battery backup for things like monitors).
 
Yeah don't know why that Bill guy is actively fighting pure sine wave UPSes. One is obviously better but costs more due to the parts they have to put in to the UPS to make it pure sine wave and the stepped sine wave UPS is cheaper and requires less parts to make etc. One is good enough and the other is obviously better, if you afford the better obviously buy the better one.

Strive to be better, not just be good enough.
 
Yeah don't know why that Bill guy is actively fighting pure sine wave UPSes.
I am NOT actively fighting them. I am actively fighting the nonsense spewed by that Gasaraki guy and others like him who claim pure sine wave UPS are "needed" with todays computers. That is just pure marketing hogwash.

Strive to be better, not just be good enough.
:( If folks like that Gasaraki guy would just read what I actually said multiple times, they would see where I said to get a "good" UPS with AVR. I never said to get one that is just "good enough". They would see where I posted links to multiple sources (EVGA, Seasonic for example) that said their supplies work fine with simulated sinewave UPS.

AND, folks like that Gasaraki guy would also have noted that I clearly said,
I am NOT saying to avoid pure sinewave. If you can find one in your budget that meets your needs, go for it. But do not discount an otherwise fine UPS just because it outputs (when on battery) a stepped or approximated sinewave.

So what I am actively fighting, Gasaraki, is folks like you who spread misinformation about what others say and about products they clearly know little about.
 
I asked the guy who wrote this book

Switching Power Supply Design.jpg


and showed him the folowing
and he felt that in this case the supply (SilverStone OP700) was badly designed to have reacted so badly to square wave UPS units.

So while I agree that a square wave UPS should be fine, there are some times where this may not be the case.
 
Last edited:
I asked the guy who wrote this book

View attachment 382374

and showed him the attached graphs and he felt that in these cases the supply was badly designed to react so badly to square wave UPS units.

So while I agree that a square wave UPS should be fine, there are some times where this may not be the case.

So this is not so clear cut, because with any square wave UPS there is a question of how fast the square wave transition occurs (risetime/falltime). In a switching power supply there is typically a rectifying circuit and a capacitor to provide internal DC bus. In some cases, the power supplies could be designed with a less powerful component that is designed to switch at zero crossing - a strategy that works well for sine-wave inputs, but that can run into trouble when presented with square wave input. You could also have a feedback happening between UPS output and power supply input that could result in out-of-spec voltage (for example, overshoots during transition). Personally, if the power supply was not tested with square wave input I would not assume it can handle it without a look at the schematic first. A stepped sine-wave should be better..
 
Last edited:
One more consideration - a fairly common way to feed an auxiliary circuit that needs low voltage is to pass input AC through a small capacitor. The capacitor acts as a resistance and the voltage is lowered. However, the way the capacitors work is that this resistance depends on input frequency. A sharp square wave has a lot of higher harmonics that could fry a circuit that was designed to work fine for a range of 50-60 Hz.

Perhaps the SilverStone OP700 is the only supply so badly designed and despite the high voltage on the primary cap, it would seem the supply survived.
Could also be that its capacitors partially dried out and it was not able to handle square wave.
 
:( Come, on Shrek. Every time you do this. Why? Obviously to obfuscate the issue with misinformation. Yes, misinformation. :(
in these cases the supply was badly designed to react so badly to square wave UPS units.
Where has anyone on this site ever suggested anything but quality power supplies?

And perhaps most importantly, NOT A SINGLE UPS designed specifically for computers outputs "square" waves. And you know that, Shrek!

A simulated, approximated, or modified sinewave is NOT a square wave. So PLEASE Shrek. Stop with this nonsensical misinformation.

Even Eaton clearly states, as I showed in post #15 above, that modified sine wave UPS are fine with computer systems.

This is a square wave
1738179213254.png


This is a modified sinewave
1738179331297.png


See the difference?

Let me remind you all again that entire towns in many communities use DC and inverters to power homes. Boats, RVs, travel trailers and more use batteries and inverters to power everything - have for decades - with NO damage or ill effects to their connected, and often VERY SOPHISTICATED electronics using switching power supplies.
 
First, kudos to you for wanting to protect your sensitive electronics with a good UPS with AVR. :) IMO, every computer should be protected by a good UPS.
This is one thing tech youtubers need to drill into people that are less savvy with computers especially when rating new setups. Next to no one is ever on a UPS.
 
Last edited:
It is square waves that are the issue
NO THEY AREN'T! The only reason they become an issue is because YOU, Shrek, keep bringing them up. :( Over and over and over again.

NO ONE with any inkling of sense and an ounce of experience would EVER recommend a square (or rectangular) wave UPS to be used on a computer system. NO APC, CyberPower, Tripplite, Eaton UPS marketed for computer use outputs square waves.

So why do you keep mentioning them - except to stir up confusion? :(

Yes, there are square/rectangular UPSs. But they ARE NOT for computers!!! They are to power things like emergency lights and emergency EXIT signs so people can find their way out of buildings during a full power outage.

Square wave UPS are NOT for computers. NOT for network gear. NOT for home theater audio and video equipment. NOT for life support medical equipment. NOT for air traffic control radio, navaids and instrument landing systems.

Simulated UPS have been used for all those critical systems for decades!

NOBODY (but you!) is talking about poor quality PSUs and square wave UPS. ALL the good advice given on this site is for quality PSUs from reputable makers, and "good" UPS with AVR.

"The least desirable output waveform type, a square wave"
See? You are still harping about square waves. :( Why? It makes no sense, Shrek.
 
I like this diagnosis :

This is a USB port error/warning ( not very unusual, a lot of USB connections try to draw more power than the system would like. A UPS usually does not need to draw power). Is the UPS connected via USB? If it is, is the error eliminated by unplugging the USB connection?

Odds are your ZFS drive is drawing too much power. There are 'helper USB cable that connect to (2) USB ports to supply the necessary power.

https://www.amazon.com/ALINNA-Female-Extension-Splitter-Adapter/dp/B098L4HTVG

https://www.amazon.com/CY-Type-C-Power-USB3-0-USB2-0/dp/B0C1WP8MMS?th=1
 
NO APC, CyberPower, Tripplite, Eaton UPS marketed for computer use outputs square waves.

"Centralion model 600VA 800VA 1000VA 2000VA"
"APC model 650"

Quote models and the oscillscope outputs confirm this.

Unfortunately some people call a rectangular wave 'modified sine'

inverter_waveform.jpg


So we agree square/rectangular wave UPS units are not good for computers?
 
Last edited:
  • Like
Reactions: qxp
Okay fine, Shrek.

Even though specific model numbers are not specified, all of a sudden,

after decades and decades of no problems,
after more than 50 years with switching power supplies in sophisticated electronics with no problems,
after ATX Form Factor switching PSUs have been around since the 1990s with no problems,
after "active" PFC power supplies have been around since the 1980s (yes, 10 years before ATX switching),
after EVGA, Seasonic, EATON and other major players in the power supply industry report no problems,
even after you yourself claim modified sinewave are fine,

Your one source trumps EVGA, Seasonic, EATON, and others, and

now, all of a sudden, even when the output of a UPS is only being generated during a full power outage - essentially less than .1% of the time, if that,

everyone needs to worry about square wave UPS because, according to your one source, they are such a huge problem. Got it.

Oh, and how do we know they have become such a huge problem? Obviously by the millions and millions computer, router, modem, AV equipment, ATC radios, big screen TVs, and other suddenly destroyed power supplies out there that have been running fine for years with approximated sinewave UPS. :kookoo:
 
NO ONE with any inkling of sense and an ounce of experience would EVER recommend a square (or rectangular) wave UPS to be used on a computer system.

We agree!
 
So we agree square/rectangular wave UPS units are not good for computers?
That was never in dispute. Hence the frustration over it constantly brought it up any time the subject of pure vs approximated comes up.
 
Then all is fine and let us recommend against square/rectangular wave UPS units.
 
As I noted, I always (or at least try to always) specify "good" UPS with AVR.

I don't know how a potential buyer would know how to spot and avoid a square wave, unless the box said "square wave" output on it. I think their best bet is to simply avoid the cheapest unit on the shelf and to make sure it says approximated (or stepped, simulated or modified) sinewave with AVR, or pure sinewave with AVR.

If it just says battery backup (and not "uninterruptible" power supply), it should be avoided as that is a basic battery backup for non-critical, unsophisticated systems like emergency lighting, emergency EXIT signs. They would even be fine for most security cameras. But not sensitive computer electronics.

And just to reiterate again for "that guy" and others like him who doesn't read what others say, I am NOT against pure sinewave UPS. If you can find one in your budget that meets your needs, go for it. But do not discount an otherwise fine UPS just because it outputs (when on battery) a stepped or approximated sinewave.
 
A most valid point, how does one know?
 
how does one know?
As noted and showed above in post #83, there is clearly a difference between a square wave and a modified, stepped, or approximated waveform. Quite certain there would be some serious legal liability issues if a UPS maker advertised a square wave UPS as something other than a square wave. I guess just another reason to stick with known brands like APC, CyberPower, Eaton, Tripp-lite (now owned by Eaton), and avoid the budget models.
 
I've attached the USB disks to my USB hub and now the kernel log is more clean,except for one error that's still there and that certify that I'm affected by the classic X ; Y problem :

Code:
[ 8344.524198] WebExtensions[4280]: segfault at 55fb19c22f3c ip 000075fb1fb1d2f3 sp 00007ffedb7b0a90 error 4 in libxul.so[75fb1af46000+
6526000] likely on CPU 12 (core 4, socket 0)

[ 8344.524206] Code: 00 e9 b1 c8 ff ff 48 8b 74 24 18 48 8b 06 49 8b 0b 48 c7 04 c1 03 00 00 00 48 8b 06 49 8b 0b 48 89 6c c1 08 48 83
06 02 eb c5 <41> 8b 55 f4 41 be 01 00 00 00 85 c9 74 b7 85 d2 74 9a 48 8b 74 24

[20836.485594] WebExtensions[9956]: segfault at 55faca86fcd0 ip 000075fb1fb1d54b sp 00007ffedb7b72d0 error 4 in libxul.so[75fb1af46000+
6526000] likely on CPU 3 (core 3, socket 0)

[20836.485603] Code: e9 0a fe ff ff 66 2e 0f 1f 84 00 00 00 00 00 4d 39 fc 0f 83 47 c6 ff ff 49 8b 02 48 ff c8 49 89 02 48 85 c0 0f 8e
73 02 00 00 <4b> 8b 6c e5 00 49 ff c4 48 b8 ff ff ff ff ff ff fa ff 48 39 c5 76

[21732.688534] WebExtensions[17063]: segfault at 6988f100000 ip 000075fb1fb04fc5 sp 00007ffedb7b2740 error 4 in libxul.so[75fb1af46000+
6526000] likely on CPU 4 (core 4, socket 0)

[21732.688547] Code: 10 bd 8c 00 00 00 45 31 e4 4a 8b 5c e6 08 48 83 fb 02 74 43 48 89 d8 48 83 e0 07 0f 85 d7 00 00 00 48 89 d8 48 25
00 00 f0 ff <48> 83 38 00 75 27 89 d9 c1 e9 03 89 da 41 b8 01 00 00 00 49 d3 e0

So I go for a bug in Firefox. What do you think brothers ?
 
Full dmesg, you might not be seeing the right error. The error that you posted shows segfault in libxul.so - this is related to the user interface, and is probably an induced error. Since Firefox runs fine for so many people, I would first update to the latest stable version and remove any addons or extensions. If the error is still there, it is likely caused by graphics card driver or the graphics card itself.
 
I saw the bug reported more times in the recent past,anyway. This is the full kernel log :


I don't see any other error that can induce the Firefox one.

Do you like this error ?

Code:
[50626.815437] traps: Isolated Web Co[94124] general protection fault ip:75fb1f3846f2 sp:7ffedb7b41d0 error:0 in _libxul.so (deleted)[75fb1af46000+6526000]

the error is reported even if I have removed and reinstalled Firefox...the file is not on the disk anymore and perhaps its name is changed,but...it crashes anyway :D
 
Looking at the full dmesg the first thing off that I see is this:

34539.429634] ------------[ cut here ]------------
[34539.429635] UBSAN: shift-out-of-bounds in /build/linux-vCyKs5/linux-6.8.0/fs/ufs/super.c:1247:19
[34539.429637] shift exponent 36 is too large for 32-bit type 'int'
[34539.429639] CPU: 7 PID: 29594 Comm: mount Tainted: P O 6.8.0-51-generic #52-Ubuntu
[34539.429641] Hardware name: Gigabyte Technology Co., Ltd. Z390 AORUS PRO/Z390 AORUS PRO-CF, BIOS F12g GA9 06/08/2020
[34539.429642] Call Trace:
[34539.429644] <TASK>
[34539.429646] dump_stack_lvl+0x76/0xa0
[34539.429649] dump_stack+0x10/0x20
[34539.429650] __ubsan_handle_shift_out_of_bounds+0x199/0x370
[34539.429654] ufs_fill_super.cold+0x16/0x1b7 [ufs]
[34539.429658] ? sb_set_blocksize+0x1d/0x70
[34539.429682] ? __pfx_ufs_fill_super+0x10/0x10 [ufs]
[34539.429685] mount_bdev+0xf3/0x140
[34539.429688] ufs_mount+0x15/0x30 [ufs]
[34539.429711] legacy_get_tree+0x28/0x60
[34539.429714] vfs_get_tree+0x27/0x100
[34539.429716] do_new_mount+0x1a0/0x340
[34539.429718] path_mount+0x1e0/0x830
[34539.429719] ? putname+0x5b/0x80
[34539.429721] __x64_sys_mount+0x127/0x160
[34539.429723] x64_sys_call+0x1e57/0x25a0
[34539.429725] do_syscall_64+0x7f/0x180
[34539.429728] ? put_fs_context+0xff/0x1c0
[34539.429730] ? do_new_mount+0x10d/0x340
[34539.429731] ? rseq_get_rseq_cs+0x22/0x280
[34539.429753] ? rseq_ip_fixup+0x90/0x1f0
[34539.429755] ? syscall_exit_to_user_mode+0x86/0x260
[34539.429757] ? do_syscall_64+0x8c/0x180
[34539.429758] ? do_faccessat+0x1c2/0x2f0
[34539.429780] ? syscall_exit_to_user_mode+0x86/0x260
[34539.429782] ? do_syscall_64+0x8c/0x180
[34539.429783] ? __do_sys_newfstatat+0x53/0x90
[34539.429786] ? syscall_exit_to_user_mode+0x86/0x260
[34539.429787] ? do_syscall_64+0x8c/0x180
[34539.429807] ? syscall_exit_to_user_mode+0x86/0x260
[34539.429809] ? do_syscall_64+0x8c/0x180
[34539.429810] ? irqentry_exit_to_user_mode+0x7b/0x260
[34539.429812] ? irqentry_exit+0x43/0x50
[34539.429813] ? clear_bhb_loop+0x15/0x70
[34539.429815] ? clear_bhb_loop+0x15/0x70
[34539.429816] ? clear_bhb_loop+0x15/0x70
[34539.429817] entry_SYSCALL_64_after_hwframe+0x78/0x80
[34539.429819] RIP: 0033:0x7689eff2af0e
[34539.429838] Code: 48 8b 0d 0d 7f 0d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d da 7e 0d 00 f7 d8 64 89 01 48
[34539.429839] RSP: 002b:00007ffd8f391668 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
[34539.429841] RAX: ffffffffffffffda RBX: 0000575a9b52abe0 RCX: 00007689eff2af0e
[34539.429842] RDX: 0000575a9b52b070 RSI: 0000575a9b52b0b0 RDI: 0000575a9b52b090
[34539.429843] RBP: 00007ffd8f3916d0 R08: 0000575a9b52be30 R09: 0000000000000007
[34539.429844] R10: 0000000000000001 R11: 0000000000000246 R12: 0000575a9b52b090
[34539.429845] R13: 0000575a9b52b0b0 R14: 0000575a9b52b070 R15: 0000575a9b52ad40
[34539.429846] </TASK>
[34539.429847] ---[ end trace ]---

There is a launchpad bug for this, so people are working to fix it.

This is followed immediately by

[35018.100057] ufs: error (device sdc2): ufs_check_page: bad entry in directory #19957332: rec_len is smaller than minimal - offset=512, rec_len=0, name_len=0
[35018.100082] ufs: error (device sdc2): ufs_readdir: bad page in #19957332

which indicates that your ufs has corruption, or that the bug above induces appearance of corruption.

Do you have a UFS partition that you are trying to mount ? Try to boot without mounting it.

Launchpad bug link:

 
Back
Top