|
Every computer needs ECC for main memory, it's just that PC clone companies of the 80s and 90s stripped the feature for consumers in order to save a few bucks - so most people who use computers have never experienced what it's like when a system doesn't just periodically throw errors. ZFS does better than other filesystems without ECC, but it's not perfect - even if you enable in-memory checksumming (which fucks with performance a lot), it's still always better to have ECC than not to have it.
|
#
?
Jul 16, 2023 11:19
|
|
|
|
| # ? May 9, 2024 23:37 |
|
|
I heard an unusual noise from my server that has a ZFS RaidZ2 array with 5x8TB drives that are a few years old. Checked the status and sure enough one of the drives was reporting errors. I shuffled some data around and swapped in another 8TB I'd had sitting in my main PC, re-silvered and all looks fine. I pulled the bad drive and stuck it in my main PC, moved some data to it as a test and everything looks fine. CrystalDiskInfo shows no issues, etc. I'm not planning on ever trusting it with anything critical, but if it's not dead I wouldn't mind throwing it inside an enclosure and using it as an additional offsite backup drive. Is there any stress tests I can do to try to validate that the drive is indeed 'healthy', or at least as healthy as a several-years-old drive can be?
|
|
#
?
Jul 17, 2023 05:27
|
|
|
apologies because I think I'm missing something obvious, but ever since a win11 update I've had to remap the drives from my server (nothing special, synology DS418play) every time I restart my computer after disconnecting the old mapped drives. trying to connect has windows throw a local device is still in use error, but it persists even after rebooting the NAS. had the same setup until a month ago and never ran into the issue before, was there a DSM update tightening security?
|
|
#
?
Jul 18, 2023 07:41
|
|
|
If the error message actually is "the local device name is already in use", then that looks some minor thing got borked in windows. Like even when it wasn't restored properly for whatever reason, a mapping for a share with that same name already exists, is what it's complaining about while you're trying to map it again. The most common suggestion I see is to wipe the whole list of mapped shares with net use * /delete from the command line with admin privileges and then remap. Presumably it would suffice to just delete the one you're having problems with if there are more. Other thing to look at is whether file and printer sharing is still turned on and they're also saying you could get this if there is not enough free space on the share (but they don't qualify this). In any case it doesn't look to be a symptom of heightened security on DSM's end.
|
|
#
?
Jul 18, 2023 10:31
|
|
|
Another way to bypass it all is to access everything via UNC paths. I have all various shares and subfolders thereof pinned as UNC paths in Explorer's quick access section. There's rarely any apps that don't support these nowadays. So long you logged in once and told it to save credentials, it's all smooth.
|
|
#
?
Jul 18, 2023 10:49
|
|
|
got it, thanks!
|
|
#
?
Jul 18, 2023 10:58
|
|
|
Hey there again. If you recall from some posts a few pages back I am on a disk shelf installation adventure and still having problems. I summed it up in this Reddit post https://www.reddit.com/r/homelab/comments/152mb2b/issues_connecting_netapp_ds4246_with_truenas_scale/ and I wanted to cross post it here to see if anyone can offer any insight as well. Would be a huge W if y'all could give it a quick read. I tried to be thorough as possible. I feel like I am taking crazy pills. I know with 1000% certainty that my 9305-24i (main internal SAS controller for the 24 internal bays) is working, and uses the same connection type. It has been chugging away with no problems, running ZFS/IT firmware for the better part of a year. All 24 internal HDDs, which are the same model as the ones in the shelves, are working as intended on that card. So I am going to try and just run the cables from the shelf to this known-to-be-working card instead of the new 9305-16e and see if they get detected. At least that will likely pinpoint if I am running into an issue with the shelf or the card. I am kind of praying it is the HBA card. Because I don't even know where to begin in terms of ripping apart the shelf itself to test there. Someone in the reddit thread suggested checking for interposers. The shelf is from eBay so I don't know maybe there is a miniscule chance I got two shelves with 48 interposers for SATA and it was never on the listing, and I never felt any problems on a physical level connecting the disks in the bays. Any other suggestions would be so very helpful.
|
|
#
?
Jul 18, 2023 18:10
|
|
|
Do you see the enclosure at all when you run lsscsi?
|
|
#
?
Jul 18, 2023 21:29
|
|
|
IOwnCalculus posted:Do you see the enclosure at all when you run lsscsi? Nope. Card, yes. Nothing to indicate the enclosure itself.
|
|
#
?
Jul 18, 2023 23:04
|
|
|
UnRaid 6.11: Is there a better way to back up the cache drive than to use CA Backup / Restore Appdata? I just had my cache drive fried by a power flicker (time to get a UPS...) and while I was fortunate to have a recent backup I still had to spend a couple hours troubleshooting and fixing SQL DB failures for Plex and the *arr apps.
|
|
#
?
Jul 19, 2023 04:04
|
|
|
Windows 98 posted:Nope. Card, yes. Nothing to indicate the enclosure itself. Do you have both power supplies actually powered or are you trying to run it on one? Green LEDs on both power supplies? Have you tried unplugging the second IOM from the midplane and booting the DS4246 with just one IOM installed? I was having some strange behavior out of my DS4246 the other day and the only thing that brought it back was going to an absolute minimum config and hot-adding the rest. One PSU, one (Xyratex) IOM, no drives plugged in. Then add the other PSU. Then see if it shows on lsscsi (possibly after running rescan-scsi-bus but it shouldn't need to).
|
|
#
?
Jul 19, 2023 04:34
|
|
|
IOwnCalculus posted:Do you have both power supplies actually powered or are you trying to run it on one? Green LEDs on both power supplies? I am using both. I have done experiments with just one PSU running but holy gently caress is it loud so I stopped doing it haha. I have not tried unplugging the IOMB. I will give it a shot,!
|
|
#
?
Jul 19, 2023 04:36
|
|
|
Tiny Timbs posted:UnRaid 6.11: Is there a better way to back up the cache drive than to use CA Backup / Restore Appdata? I just had my cache drive fried by a power flicker (time to get a UPS...) and while I was fortunate to have a recent backup I still had to spend a couple hours troubleshooting and fixing SQL DB failures for Plex and the *arr apps. I don't know of anything else for the entire cache drive, but Sonarr and Radarr have their own backup system baked in. I have their databases back up to my Google drive in addition to the CA Backup/Restore plugin.
|
|
#
?
Jul 19, 2023 12:16
|
|
|
Scruff McGruff posted:I don't know of anything else for the entire cache drive, but Sonarr and Radarr have their own backup system baked in. I have their databases back up to my Google drive in addition to the CA Backup/Restore plugin. Oh cool, I didn’t know that. I’ll get that set up then. Thanks!
|
|
#
?
Jul 19, 2023 12:44
|
|
|
CA Backup/Restore is supposed to be one of the more reliable ways to get a backup of your appdata because it shuts down your docker containers before making its backup. Did you put the files back in place manually or use the restore function of the plugin? I've had permissions issues manually copying files from its backup to a single borked application, but for a full cache drive replacement you should just need to hit restore and have everything copied back correctly. Spaceinvader one is putting out video tutorialss with ZFS on Unraid 6.12, and one of the things you can do with it is setup each folder in your appdata as an individual zfs dataset, and then use zfs replication to backup those to a zfs formatted drive in the array. But his last video on actually setting up the replication isn't out yet.
|
|
#
?
Jul 19, 2023 13:48
|
|
|
Make sure you have the latest app data plug-in, it was forked a bit ago
|
|
#
?
Jul 19, 2023 16:36
|
|
|
Computer viking posted:And yes, of course ECC memory is useful in anything where data you want to keep does at any point reside in memory. The reason people are touchy when people talk about ZFS+ECC is that one old article that predicted that bad memory would lead ZFS to self-corrupt into nothingness worse than other file systems, which seems to be based on a misunderstanding. Anecdote, but I recently had a stick of ram fail on a system running ZFS and the file system survived just fine, even if my uptime did not. Took me about 4 or 5 days to root cause the issue. Is there a better way to monitor for memory related errors so I can get more warning in the future?
|
|
#
?
Jul 19, 2023 17:26
|
|
|
Talorat posted:Anecdote, but I recently had a stick of ram fail on a system running ZFS and the file system survived just fine, even if my uptime did not. Took me about 4 or 5 days to root cause the issue. Is there a better way to monitor for memory related errors so I can get more warning in the future? One of the things people (nominally) pay for when purchasing server hardware, is the ability to know for sure that the firmware will generate Machine Check Exceptions (a form of Non-Maskable Interrupt, similar to how a PC will beep if you hold down 6 keyboard keys on a keyboard without n-key rollover). It's then up to the OS to properly file this in the logging system, and you'll usually want something to decode it too, since the format is old and bonkers. So unless you can verify for sure that your machine gets these by using a known-bad DIMM, Rowhammer, or some other way - there's no real way of knowing. Unfortunately, it's not exactly uncommon for server vendors to promise that they can do this, and then just utterly fail to deliver. I have an anecdote that I think I've mentioned before ITT, where in the course of buying systems for two datacenters for an active-active failover cluster over an extended period of time, one datacenter would regularly keel over for reasons that couldn't be explained - and we only accidentally found out about it being an undocumented (as in no Product Change Notification) revision from the vendor. And I've heard similar anecdotes from other people, including one person who's now working on building an entirely new rack-scale server design, similar to how the hyperscalers build theirs. BlankSystemDaemon fucked around with this message at 17:36 on Jul 19, 2023 |
|
#
?
Jul 19, 2023 17:32
|
|
|
Anyone got a UPS recommendation? Just had a wicked storm blow through and flicker the power a few times and real quick succession.....and it reminds me that my NAS isn't well protected beyond a surge protector. It's a Synology 720+. Don't need anything all that powerful.
|
|
#
?
Jul 20, 2023 00:47
|
|
|
RestingB1tchFace posted:Anyone got a UPS recommendation? Just had a wicked storm blow through and flicker the power a few times and real quick succession.....and it reminds me that my NAS isn't well protected beyond a surge protector. It's a Synology 720+. Don't need anything all that powerful. I use an Eaton 3S 550. It also supports a Network UPS Tool server so it can communicate with your NAS and other devices so they can safely shut down when the UPS is running low on power.
|
|
#
?
Jul 20, 2023 00:56
|
|
|
RestingB1tchFace posted:Anyone got a UPS recommendation? Just had a wicked storm blow through and flicker the power a few times and real quick succession.....and it reminds me that my NAS isn't well protected beyond a surge protector. It's a Synology 720+. Don't need anything all that powerful. I've got a lot of cyberpowers that are fine, but with their recent issues I'd suggest APC. They have a higher price tag but I have a APC Back UPS 1000 and 1500 and they've both been fine. You probably won't need something that big if it's just your UPS but I have a 1000VA for my NAS, router, wireless access point and a big switch. It's overkill but it keeps the infrastructure alive longer in an outage so I can finish what I'm doing and sign out or whatever. This also gives me a central point to plug in when I fire up the router and access point with generator power when I want to get online during an outage with my laptop. I leave the NAS off if it's like that.
|
|
#
?
Jul 20, 2023 04:55
|
|
|
RestingB1tchFace posted:Anyone got a UPS recommendation? Just had a wicked storm blow through and flicker the power a few times and real quick succession.....and it reminds me that my NAS isn't well protected beyond a surge protector. It's a Synology 720+. Don't need anything all that powerful. I use this APC one which has consistently been worked great for 3 years now and has never failed to kick in during an outage. It has enough power to run my server for about 20-30 minutes (actually maybe more now I haven’t tested) and enough juice to keep my networking hardware on for 2-3 hours if I power down the server.
|
|
#
?
Jul 20, 2023 07:51
|
|
|
Talorat posted:I use this APC one which has consistently been worked great for 3 years now and has never failed to kick in during an outage. It has enough power to run my server for about 20-30 minutes (actually maybe more now I haven’t tested) and enough juice to keep my networking hardware on for 2-3 hours if I power down the server. I also have this one and it's worked well for the couple months I've had it. Bought it intending to run my main PC off it, but it has my nas box and network gear at 22% load.
|
|
#
?
Jul 20, 2023 13:47
|
|
|
Windows 98 posted:Hey there again. If you recall from some posts a few pages back I am on a disk shelf installation adventure and still having problems. I summed it up in this Reddit post https://www.reddit.com/r/homelab/comments/152mb2b/issues_connecting_netapp_ds4246_with_truenas_scale/ and I wanted to cross post it here to see if anyone can offer any insight as well. Would be a huge W if y'all could give it a quick read. I tried to be thorough as possible. I feel like I am taking crazy pills. ok so I already had a SAS controller (internal) for the original 24 drives in the system. It's same SAS connector type as the new HBA. So I took the new card out, disconnected all my drives, and then connected the shelf direct to the internal card. I know the card works. I've been using it for my 24 internal drives for 8 months. Nothing. No drives, no shelf shows up in any capacity with CLI commands. SAS card still showing up like normal though. So I am pretty sure it's something wrong with the shelf and not the new HBA. how the gently caress am I supposed to troubleshoot that. there is no OS on a shelf. there is no way to remote into a shelf. there is no video output for a monitor on a shelf.
|
|
#
?
Jul 20, 2023 23:26
|
|
|
Nitrousoxide posted:I use an Eaton 3S 550. It also supports a Network UPS Tool server so it can communicate with your NAS and other devices so they can safely shut down when the UPS is running low on power. That looks like it would fit my needs. Like the idea that it would be able to shut down the NAS properly as power was running low. Some of the others are way more than I need. This is specifically to protect my NAS......which is used almost entirely for storage. I do run a personal Plex server off it.....but I'd be home when using it.....and I don't use it very often....so I'm not too concerned about it being in use during a power outage when I wasn't around. Now.....if anyone has a good reason why I should buy something larger.....I'm all ears. Note that I'll probably never be running a full fledged server or desktop anytime soon.
|
|
#
?
Jul 21, 2023 01:05
|
|
|
RestingB1tchFace posted:That looks like it would fit my needs. Like the idea that it would be able to shut down the NAS properly as power was running low. To setup your Synology NAS as the power server: https://www.reddit.com/r/synology/comments/gtkjam/use_synology_nas_as_ups_server_to_safely_power/
|
|
#
?
Jul 21, 2023 01:08
|
|
|
I'm expanding the storage in my Synology 418 and I'm wondering about something: I had 3 disks of 3TB each, in Raid5. I replaced each disk with a new 8TB disk, one at a time. I had an empty bay so I used the replace function: insert new disk, tell Synology to replace one of the 3TB disks with the new one, and let it do its thing. Took like 6 hours to copy all the data over, after which it deactivated the old drive. Repeat 3 times and now there are 3 8TB disks. (And a deactivated 3TB one.) After the last one, it then automatically started expanding the array into the empty space. I expected this to take like 5 seconds because all it should be doing is adding empty space, but it has already been several hours and judging from the progress it will take, I guess, upwards of 10 hours total? Any idea what the hell it's doing? Why is it taking so long to add a whole lot of emptiness?
|
|
#
?
Jul 21, 2023 01:21
|
|
|
uXs posted:Any idea what the hell it's doing? Why is it taking so long to add a whole lot of emptiness? the way I understand it hard drives never actually "delete" any data. it just keeps moving through the sectors and eventually may circle back around and replace some of the nonsense data once there is no more actual free spots for data to go. It's probably converting a ton of 1s into 0s and actually deleting it. just a theory. im probably wrong.
|
|
#
?
Jul 21, 2023 01:51
|
|
|
If the raid is done in a way that isn't file system aware maybe it has to sync the drives up?
|
|
#
?
Jul 21, 2023 04:14
|
|
|
Rexxed posted:I've got a lot of cyberpowers that are fine, but with their recent issues I have a Cyberpower 1500 on my desktop; what issues are these? I haven’t heard anything but I joined the thread recently and only read the first/last four pages. I’m on vacation right now, and the only things left turned on in my room is my T-Mobile base station and Cyberpower UPS.
|
|
#
?
Jul 21, 2023 17:24
|
|
|
DerekSmartymans posted:I have a Cyberpower 1500 on my desktop; what issues are these? I haven’t heard anything but I joined the thread recently and only read the first/last four pages. I’m on vacation right now, and the only things left turned on in my room is my T-Mobile base station and Cyberpower UPS. A year or two ago some CyberPowers started spontaneously combusting. Apparently they used either the wrong glue or the wrong amount of glue on a certain connection inside the unit, and it would short and catch fire. I don't think there was ever a recall on them?
|
|
#
?
Jul 21, 2023 18:05
|
|
|
Cute as my Fractal Node NAS is, the 12 drives I've janked into it are getting a bit warm (50'c under load). A bit of vodka braining and a visit to AliExpress a few weeks back meant that 6 120mmx38mm Delta case fans have arrived with a claimed 210CFM and 21mm of static pressure which should be just the ticket. Might have to look at running CAT6 to the garage if they are half as geneva convention breakingly noisy as I remember them being in the early 2000s 
|
|
#
?
Jul 21, 2023 18:48
|
|
|
Windows 98 posted:ok so I already had a SAS controller (internal) for the original 24 drives in the system. It's same SAS connector type as the new HBA. So I took the new card out, disconnected all my drives, and then connected the shelf direct to the internal card. I know the card works. I've been using it for my 24 internal drives for 8 months. There's a serial interface on the IOM6, you will probably need a custom cable of some sort to use it. The serial interface on my Xyratex controller is exposed via a 3.5mm TRS connector 
|
|
#
?
Jul 21, 2023 20:10
|
|
|
Wizard of the Deep posted:A year or two ago some CyberPowers started spontaneously combusting. Apparently they used either the wrong glue or the wrong amount of glue on a certain connection inside the unit, and it would short and catch fire. I don't think there was ever a recall on them? Yeah, it seemed like an issue with fairly new units at the time so I figure my older ones are probably fine since I've had most of them 5-15 years at this point, but I've bought a couple of APCs in the last three years or so. The only thing I don't like about APC is that APC seems to be stuck on the idea of selling replacement batteries as a cartridge which is just two normal 12V SLA batteries taped together to a connector that connects to the unit as a plug. That way they can sell you $40 (each one costs about $20) of batteries for $80. You can just buy two batteries and replace them on the connector and put some packing tape on the sides as a replacement. It's one extra step that cyberpower didn't bother to upcharge for, although they'll sell you cyberpower approved batteries for double the normal price as well. I just get 8x mighty max 12V 9Ah for $160 on amazon or whatever every two or three years since I have a lot of UPSes.
|
|
#
?
Jul 22, 2023 01:46
|
|
|
Wizard of the Deep posted:A year or two ago some CyberPowers started spontaneously combusting. Apparently they used either the wrong glue or the wrong amount of glue on a certain connection inside the unit, and it would short and catch fire. I don't think there was ever a recall on them? Yikes! I’ll look into it rn, but I’m not going to be home till Sunday (if home is still there!  ).
|
|
#
?
Jul 22, 2023 07:31
|
|
|
Rexxed posted:Yeah, it seemed like an issue with fairly new units at the time so I figure my older ones are probably fine since I've had most of them 5-15 years at this point, but I've bought a couple of APCs in the last three years or so. The only thing I don't like about APC is that APC seems to be stuck on the idea of selling replacement batteries as a cartridge which is just two normal 12V SLA batteries taped together to a connector that connects to the unit as a plug. That way they can sell you $40 (each one costs about $20) of batteries for $80. You can just buy two batteries and replace them on the connector and put some packing tape on the sides as a replacement. It's one extra step that cyberpower didn't bother to upcharge for, although they'll sell you cyberpower approved batteries for double the normal price as well. I just get 8x mighty max 12V 9Ah for $160 on amazon or whatever every two or three years since I have a lot of UPSes. that's the same thing? or you mean that cyberpower doesn't bother with the double-sided tape on the batteries? I had a business account at batteries+ that brought their prices down close to amazon level and every time I replaced APC they happily taped them together and transferred the connectors for me. It was worth it since A) same day replacement and B) I'd have to drop off the old ones for recycling anyway.
|
|
#
?
Jul 23, 2023 19:46
|
|
|
Harik posted:that's the same thing? or you mean that cyberpower doesn't bother with the double-sided tape on the batteries? Cyberpower will sell you two batteries as a kit but just has instructions to replace them. APC puts them together taped up with an adapter and a special cartridge plug (which you can just reuse). Not exactly the same but yes, they both have high pricing because that's free money for them.
|
|
#
?
Jul 23, 2023 22:26
|
|
|
There really is no ATX non-rack mount mini ITX NAS case out there that does 16 drives, is there? The Fractal Design Node 804 in theory could handle it if there were more 3.5" bay mount points available but shoving that many drives into as small of a case as possible (with adequate cooling, of course) seems like a problem nobody's bothering with because it's basically the wheelhouse of every other enterprise storage vendor.
|
|
#
?
Jul 24, 2023 18:26
|
|
|
necrobobsledder posted:There really is no ATX non-rack mount mini ITX NAS case out there that does 16 drives, is there? The Fractal Design Node 804 in theory could handle it if there were more 3.5" bay mount points available but shoving that many drives into as small of a case as possible (with adequate cooling, of course) seems like a problem nobody's bothering with because it's basically the wheelhouse of every other enterprise storage vendor. For that many drives, you're pretty much looking for something like the Fractal Define 7 XL. It supports 18x 3.5" drives and is mITX compatible.
|
|
#
?
Jul 24, 2023 18:39
|
|
|
|
| # ? May 9, 2024 23:37 |
|
|
Tornhelm posted:For that many drives, you're pretty much looking for something like the Fractal Define 7 XL. It supports 18x 3.5" drives and is mITX compatible. Pretty much. You can get more drives than specified into the Node 804 with some of Fractal's bracket mounts and a bit of redneck engineering, but it's a right little toaster unless you put in insanely loud, high CFM, high static pressure fans. It just gets to the point where you're forcing a poor solution to such a niche problem that it doesn't make sense.
|
|
#
?
Jul 24, 2023 18:44
|
|