[BUG] RPC error: useradd failed (exit status 12) — cannot start any new Cowork tasks

Found 1 possible duplicate issue:

1. https://github.com/anthropics/claude-code/issues/30751

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

That's happening with all of my tasks

Happening to me right now. Has been happening for the last eight hours.

cant use at all. max pro user, all new, scheduled, existing, or past..running into the error

confirm i also have this issue since about 7 days ago
cowork for new tasks does not work

Reporting from Mexico | macOS | Team + Max Pro accounts

I'm hitting the exact same error since approximately April 5–6, 2026:

RPC error: ensure user: useradd failed: exit status 12: useradd: cannot create directory /sessions/...

Details:

• Account types: Team (one account) + Max Pro (second account) — both affected
• Region: Mexico
• OS: macOS
• Claude Desktop version: 1.1348.0 (efa3fd) — built 2026-04-08T15:06:08.000Z
• Issue started: ~April 5–6, 2026
• All Cowork tasks fail (new, scheduled, existing, past) — Cowork is completely unusable
• Tried: restart, reinstall, clear cache, reset app data — error persists

Adding my +1 to confirm this is still actively happening as of April 8, 2026. Two separate accounts on different plan tiers both affected simultaneously.

I deleted and reinstalled. Sent from my iPhoneOn Apr 8, 2026, at 8:12 PM, Utopica @.***> wrote:utopica-dev left a comment (anthropics/claude-code#36465)
Reporting from Mexico | macOS | Team + Max Pro accounts
I'm hitting the exact same error since approximately April 5–6, 2026:
RPC error: ensure user: useradd failed: exit status 12: useradd: cannot create directory /sessions/...

Details:

Account types: Team (one account) + Max Pro (second account) — both affected
Region: Mexico
OS: macOS
Claude Desktop version: 1.1348.0 (efa3fd) — built 2026-04-08T15:06:08.000Z
Issue started: ~April 5–6, 2026
All Cowork tasks fail (new, scheduled, existing, past) — Cowork is completely unusable
Tried: restart, reinstall, clear cache, reset app data — error persists

Adding my +1 to confirm this is still actively happening as of April 8, 2026. Two separate accounts on different plan tiers both affected simultaneously.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

+1 — Reporting from Lithuania/UK | macOS ARM64 | Max Pro account

Same error since ~April 9–10, 2026:

RPC error: ensure user: useradd failed: exit status 12: useradd: cannot create directory /sessions/happy-friendly-clarke

• Account: Max Pro (personal) — corporate Team account doesn't have Cowork access at all
• OS: macOS Darwin 25.4.0 (ARM64)
• Claude Code version: 2.1.92
• 28 scheduled tasks disappeared — list_scheduled_tasks returns empty
• All Cowork operations fail — new tasks, scheduled, existing
• Business-critical: CRM monitoring, email triage, daily reports, backups — all down for days

This is now 7+ days for the earliest reporters. Any ETA from the team?

I have the same issue on Claude 1.2278.0 (osx) persisting for a day.

FYI following the workaround: https://github.com/anthropics/claude-code/issues/30751

Workaround

Fully quit Claude Desktop
mv "~/Library/Application
Support/Claude/vm_bundles/claudevm.bundle/sessiondata.img"
~/Desktop/sessiondata.img.bak
Restart Claude Desktop (recreates the image fresh)
Delete .bak to reclaim ~8.5GB

Works.

FYI following the workaround: #30751 Workaround Fully quit Claude Desktop mv "~/Library/Application Support/Claude/vm_bundles/claudevm.bundle/sessiondata.img" ~/Desktop/sessiondata.img.bak Restart Claude Desktop (recreates the image fresh) Delete .bak to reclaim ~8.5GB Works.

this worked for me also

thank you sir!!

i am still affected by this issue. for 3 days already. no new session cowork work and i get this error message all the time:RPC error -1: ensure user: useradd failed: exit status 12: useradd: cannot create directory /sessions/optimistic-adoring-maxwell
and RPC error -1: failed to mount /mnt/.virtiofs-root/shared/c/Users/shirl/AppData/Local/Packages/Claude_pzs8sxrjxfjjc/LocalCache/Roaming/Claude/local-agent-mode-sessions/008d4757-502a-4647-817e-d1a1160ebb85/ef1c7915-9557-4fee-8fb5-269f72b8cd1f/local_bbda9f20-0be0-41b9-9188-136507c544fa/outputs as outputs: failed to create mount point: mkdir /sessions/jolly-bold-ritchie/mnt/outputs: no space left on device

Any fix for this bug?

@syeedshah I've been able to solve it here https://github.com/msmobileapps/vm-disk-cleanup-plugin

Reproducing this on Windows 11 (Cowork research preview, desktop client, 2026-04-30) — flagging because the issue is currently labeled platform:macOS, but the error signature is identical.
Error (from a single bash call):
bash failed on resume, create, and re-resume.
resume: RPC error -1: ensure user: useradd failed: exit status 12:
useradd: cannot create directory /sessions/relaxed-beautiful-brown
create: RPC error -1: ensure user: useradd failed: exit status 12:
useradd: cannot create directory /sessions/relaxed-beautiful-brown
A single mcp__workspace__bash call attempts three recovery steps internally (resume → create → re-resume) and all three fail identically. This is not the boot-delay "Workspace still starting" message — it fails immediately on every call, every retry.
Reproducibility: 100%. Multiple sessions across the day, e.g. nifty-zen-edison, relaxed-beautiful-brown, others — only the session-name suffix changes; useradd: cannot create directory /sessions/<name> is constant.
Things that did not help:

Multiple full Cowork app restarts
Full Windows reboot
Starting a fresh session

What still works (so this is isolated to sandbox container user provisioning, not the broader Cowork/MCP layer):

File tools (Read, Write, Edit) against the workspace mount — fine
All MCP servers — connecting and responding normally

Adjacent signal (possibly related, possibly not): mid-session a <system-reminder> listed ~40 plugin-auth MCP servers as disconnected (plugin_apollo_, plugin_human-resources_slack__, plugin_productivity_clickup__, plugin_sales_close__, etc.). Suggests Anthropic-side infra instability around plugin/auth services concurrent with the sandbox provisioning failure.
Status page (status.claude.com / status.anthropic.com/history) shows no active incident matching this as of the timestamp above.
useradd exit status 12 = "can't create home directory" — looks like the per-session /sessions/<name> mount isn't being staged by the host before useradd runs, or the container's filesystem is rejecting the directory creation. Filing this on the Windows side in case it informs the macOS root-cause investigation (or vice versa).
Happy to provide additional logs, session IDs, or a console transcript if useful.

cancelling the subscription, it's not been usable for over a week.

Closing for now — inactive for too long. Please open a new issue if this is still relevant.

I had to wipe the machine I was running it on, something got corrupted, not sure what. But it killed Claude. My bet is that it was around allowing full control of the computer feature.

Second time this has happened now. I had to reinstall claude, but that clears all my history and scheduled jobs so I'd like to not keep doing this.