A long time ago I used Arq 5 for my backups. Then at some point I switched to Duplicacy, because Arq was just way too slow with its backup speed. Then Arq 6 came out which seemed to be full of issues, and so everyone who had a valid Arq 6 license was granted a free Arq 7 license, which I never used since I was happy with Duplicacy by then already…
But now recently (3 months ago), I was thinking “Well I use cloud storage where I have “unlimited” space, so why not just run two backup software simultaneously, just for some extra redundancy in case one of the software has a bug that corrupts its backup? And I still have this unused Arq 7 license I got for free, so maybe see how well it works?”. So I installed Arq 7 and set it up, because why not, it can’t hurt.
Well actually, it does hurt. In these 3 months, where my PC runs all day long, it has not managed to get anywhere close to finishing the backup. It’s still extremely slow, just like years ago when I tried Arq 5, and why I switched to Duplicacy. Arq is still so slow that it’s completely unusable in practice. On a good day it manages to upload 50 GB, but there are a lot of days for Arq that are not good days, where it spends 20 hours with just “scanning” the disk. That’s really the main problem of Arq: It seems to be completely incapable of scanning the disk in an efficient way. I think it’s because the developer of Arq is a Mac user who wrote a software that might work well on Mac, but uses the Windows API in some really inefficient way.
To be fair, my files are probably a worst-case scenario since I have many, many millions of very small files everywhere on my disk, but still, Duplicacy has no issue with those. Scanning the disk is never a significant timesink with Duplicacy, “it just works”. I still know, when I started using Duplicacy, I was impressed how I configured my backup, it started, it scanned a few minutes, and then it simply started uploading! As opposed to Arq, which I have only ever seen spending hours “trying to find files to upload” before it ever starts with any uploading.
And it’s not just the speed being super slow with Arq, the even bigger issue is that Arq seems to saturate the Windows API so much that the whole PC becomes really slow while it’s doing its scanning. It took me a while to figure out, but with Arq running, it often happens that “opening the start menu” or other super easy tasks in windows take minutes to complete. I only found out that Arq is causing it by looking in Process Monitor and seeing that Arq is doing some weird file operations tens of thousands of times per second. I don’t know why doing too many file operations grinds Windows to a halt, there’s surely also some blame to put on Windows there, but a good software would be written with that in mind and not make the PC unusable for hours just for running a backup. And it just works perfectly in Duplicacy.
It’s not just speed, but also RAM usage. I have a lot of RAM (128 GB), but it still annoyed me how Arq 7 would basically constantly use 30 GB RAM while doing its backup. I don’t mind some RAM usage, but constantly having 30 GB less just because of a backup software is just not quite fine any more… And Duplicacy needs only a few GBs when it runs.
So overall, I just wanted to write this as an appreciation for how much more efficient Duplicacy actually is compared to Arq. I’m really happy that Duplicacy exists and I don’t need to bother with something like Arq.
