Compare the time "wasted" waiting for the slower HDD vs the money "wasted" by excessive writes to the SDD. When you bill at your hourly rate for all the time waiting for the HDD, you'll probably discover that it won't take long for you to buy a new SDD by using the existing one instead of the HDD.

Multiply any wasted time by 10 to account for lost productivity due to reduced job satisfaction, context switching, disrupted focus, and similar knock-on effects caused by disk performance. The raw added wait time is minor compared to that!

If you're really concerned about uptime, buy two consumer SSDs and stick them in RAID1. If a drive dies, just replace it. This would fit in your budget. If you're concerned about data integrity, keep regular backups - this you have to do anyway.

How important, how secret (i.e. how confidential) is the data on the disk? What would happen to you if the data is lost? Can you afford some remote backup?

Why not both? Use the SSD for OS and other stuff, and put scratch or things with lots of IO writes on the HDD, if you don't care about the absolute fastest speed. You could also always RAID HDDs to make them faster and more reliable.

@Hefewe1zen, RAID HDDs might be the way to go here.

@BasileStarynkevitch, very confidential. Passwords and chat messages in the MySQL database. But, the stuff are encrypted. A remote backup wouldn't be really good for privacy in terms of this.

That should go in your question, not as a comment.

BTW, some contributors to refpersys.org are living in India and have practical experience related to your concerns. You could email them and give a lot more details, in particular explaining what will happen to you when the data is lost. Because it surely will be lost. You should explain your backup strategy, and what kind of data do you manage (military secrets, financial data, health related data, IoT related data, etc...)

@BasileStarynkevitch, I keep backups in multiple External 3 TB drives.

@FreeMan, right.

@Brian, okay...

Mount options (e.g. in /etc/fstab) are also very relevant

@BasileStarynkevitch, and I wonder why this question is so popular... I have a little bit of stage fear, so, I am scared I am gonna mess something up here. Feels so strange.

Now, obviously the SSD is way faster than the HDD. But, as we all must know: the SSD has a limited write endurance, but the HDD does not have a limited read/write endurance. - you never had a hard drive fail on you? What do you think caused that material fatigue, if not read and/or write operations?

You could use something like bcache to use the SSD as a cache for the HDD, but in my opinion you'd be better off using the HDD as one of the off-line backups for the SSD.

That Seagate HDD has specs of only 2400 power-on hours per year, and workload rate limit of 55 TB/year. There are 8760 hours in a year.

@Philipp, This question is getting so popular that I am kinda lost...

"but the HDD does not have a limited read/write endurance" - all recoding mediums have such limits. For a lot of heavily random access you might find the mechanical strain on the head movements of a traditional drive might be more wearing on it than the many writes to the SSD's flash. One of the differences between consumer and enterprise drives of both varieties is battery (or big-cap) backed or non-volatile RAM to delay and group/order writes without risk of data loss in a power cut situation.

@DavidSpillett, (facepalm), I have strikethrough'ed it.

seriously