# SysAdmin Tools

> Admin calculators in the browser: RAID, ZFS and Btrfs with mixed drives, chmod/umask, transfer, server tuning, SLA, backup retention, SSD and UPS.

Source: https://www.jpkc.com/db/en/tools/sysadmin/

## Calculate instead of guessing: seven admin calculators in one tool

How much space do six 8 TB drives really leave you in a RAIDZ2? Which `chmod` command sets a directory to exactly `2775` without leaving an old setuid bit behind? How many PHP-FPM workers fit on a server with 8 GiB of RAM, and how many snapshots will `restic forget` actually keep with your rules? Questions like these come up all the time in day-to-day admin work — and usually get answered with a rule of thumb, a half-remembered forum post or a quick back-of-the-envelope sum.

[SysAdmin Tools](https://www.jpkc.com/tools/sysadmin/) puts seven calculators for exactly these questions into one interface. Each tab solves one well-defined problem, shows its working right next to the result and, where it makes sense, hands you the finished command or config snippet to copy. The tool is aimed at **Linux admins**, **DevOps engineers**, **homelab and NAS tinkerers** and anyone who runs their own servers and would rather check the numbers than guess them.

## What SysAdmin Tools does — at a glance

- **RAID & Storage** — capacity, fault tolerance, rebuild time and rebuild risk, theoretical performance, power and drive cost for 23 layouts: classic RAID (Linux md / hardware) from JBOD to RAID 60, ZFS (stripe, mirror vdevs, RAIDZ1–3), Btrfs (raid1, raid1c3, raid1c4, raid10, raid5, raid6), plus Synology SHR-1/SHR-2 and Unraid. Mixed drive sizes are handled the way each system really handles them — ZFS with RAIDZ padding and slop space, Btrfs by simulating the chunk allocator. On top: a comparison table of all layouts and ready-made `mdadm`, `zpool` or `mkfs.btrfs` commands.
- **Permissions** — a chmod calculator with a bit grid, octal and symbolic notation, a reader for `ls -l` lines and whole `ls -la` listings, a simulator for chmod expressions such as `u=rwX,g=rX,o=`, and a umask calculator. The model is the real behavior of GNU `chmod`, including its special rules for setuid/setgid on directories.
- **Transfer** — transfer time, required rate or data amount, protocol overhead of TCP over Ethernet (MTU, IPv6, VLAN, PPPoE), a comparison table of common links from Fast Ethernet to NVMe, a TB ↔ TiB unit converter, monthly traffic, the 95th percentile for burstable billing, and the bandwidth-delay product with a suggested `sysctl` snippet.
- **Server Tuning** — a memory plan for the whole server, and from it the PHP-FPM pool, OPcache, MySQL/MariaDB (InnoDB), nginx and Apache httpd, each with a config snippet for Debian/Ubuntu or RHEL paths.
- **Availability** — SLA percentage ↔ allowed downtime, downtime ↔ availability, MTBF/MTTR, composite availability across several tiers (e.g. load balancer → web servers → database), and an error budget with burn rate.
- **Backup** — retention rules such as `--keep-daily`/`--keep-weekly` simulated with the algorithms of **restic**, **rustic** and **borg**, with a timeline, the matching `forget`/`prune` commands, a storage estimate for full, incremental, differential and deduplicated backups, the backup window and a 3-2-1-1-0 check.
- **SSD & UPS** — SSD lifetime from TBW or DWPD, converting SMART values from `smartctl` or `nvme smart-log`, UPS sizing by VA and watts, a battery runtime estimate with Peukert exponent, and power cost and heat output.

## Transparent, not a black box

The tool refuses to give you a number without showing where it came from. Under each result you can expand the formulas and assumptions (usually "Formulas & assumptions"); rules of thumb are explicitly labeled **Heuristic**, vendor figures as datasheet values. Where a result might alarm you — the URE risk when rebuilding a large RAID 5, say — the tool explains right next to it why the number is an upper bound and not a forecast.

Many of the rules have been checked against the real programs: GNU `chmod` and the `umask` of dash and bash, `restic forget --dry-run`, borg's prune code, `php-fpm`, OPcache, `httpd -t`, and MySQL and MariaDB servers. It shows in the details that rule-of-thumb calculators tend to miss — that `chmod 755` on a directory leaves a set setgid bit in place, that `df -h` rounds up, or that OPcache doesn't clamp an oversized `max_accelerated_files` value but throws it away.

Commands that can destroy data are treated accordingly: the RAID commands open with a warning, fill in missing device names with the placeholder `/dev/disk/by-id/DISK-01`, and tell you to check every drive with `lsblk` first.

## Architecture: everything in the browser

SysAdmin Tools runs **entirely client-side** in JavaScript. There is no server endpoint and no API — the PHP page only delivers the frame with the tabs; all the calculating happens in your browser. Output you paste in (an `ls -la` listing, samples for the 95th percentile, a snapshot list or the output of `smartctl`) is read locally only.

The tool remembers your inputs per tab in your browser's local storage, so everything is still there on your next visit. **Copy link** creates a link that carries the current tab's inputs in the address (in the part after `#`) — handy for sharing a calculation with colleagues. Pasted raw data such as an `ls -la` listing or `smartctl` output is deliberately kept out of both the link and storage, as are very long sample series and snapshot lists. **Reset** returns a tab to its defaults, and every tab can be reached directly via its anchor, e.g. `…/tools/sysadmin/#backup`.

## Try it now

**[→ Open SysAdmin Tools](https://www.jpkc.com/tools/sysadmin/)** — pick a tab, enter your own values, read the result along with its derivation and copy the matching command. No account, free, right in the browser.

## Related JPKCom tools

- **[Cron Job Helper](https://www.jpkc.com/db/en/tools/cron/)** — build and check the schedule for your backups and `forget`/`prune` runs as a cron expression.
- **[IP Tools](https://www.jpkc.com/db/en/tools/ip/)** — subnet and address-range calculations; deliberately not part of SysAdmin Tools.
- **[SSH Proxy](https://www.jpkc.com/db/en/tools/ssh-proxy/)** — SSH tunnels and commands for reaching the servers you size here.
- **[Docker/Podman Composer](https://www.jpkc.com/db/en/tools/docker/)** — for when the services in your tuning plan run in containers.
- **[Time Converter](https://www.jpkc.com/db/en/tools/time/)** — timestamps and time zones, e.g. for snapshot times coming from different systems.

---

There's more on the subpages: the **[manual](https://www.jpkc.com/db/en/tools/sysadmin/manual/)** with every tab and every formula, hands-on **[examples](https://www.jpkc.com/db/en/tools/sysadmin/examples/)**, and a collection of **[tips & tricks](https://www.jpkc.com/db/en/tools/sysadmin/tips/)**.

## In this section

- [Manual](https://www.jpkc.com/db/en/tools/sysadmin/manual.md): All seven tabs of SysAdmin Tools in detail: inputs, formulas, limits and generated commands for RAID, chmod, transfer, tuning, SLA, backup, SSD and UPS.
- [Examples](https://www.jpkc.com/db/en/tools/sysadmin/examples.md): SysAdmin Tools in practice: a NAS with mixed drives, RAID 6 vs. RAIDZ2, a setgid project folder, a server migration, PHP-FPM sizing and a backup plan.
- [Tips & Tricks](https://www.jpkc.com/db/en/tools/sysadmin/tips.md): Strategy and pitfalls for SysAdmin Tools: reading URE risk, RAIDZ padding, chmod on directories, sysctl, PHP-FPM, OPcache, restic vs. borg, SSD TBW.
