Varnish 6 reload

Question

I've upgraded my varnish from 6.2.x to 6.6.x. Amost everyting works Ok, but no reload.

After "start" ps show:

root       10919  0.0  0.0  18960  5288 ?        Ss   22:38   0:00 /usr/sbin/varnishd -j unix,user=vcache -F -a :80 -T localhost:6082 -f /etc/varnish/default.vcl -p thread_pools=8 -p thread_pool_min=100 -p thread_pool_max=4000 -p workspace_client=128k -p workspace_backend=128k -l 200m -S /etc/varnish/secret -s malloc,256m -s static=file,/data/varnish_storage.bin,80g

now I try to reload:

Apr  8 22:42:10 xxx varnishd[10919]: CLI telnet 127.0.0.1 5282 127.0.0.1 6082 Rd auth 0124ef9602b9e6aad2766e52755d02a0d17cd6cfe766304761d21ea058bd8b3b
Apr  8 22:42:10 xxx varnishd[10919]: CLI telnet 127.0.0.1 5282 127.0.0.1 6082 Wr 200 -----------------------------#012Varnish Cache CLI 1.0#012-----------------------------#012Linux,5.4.0-107-generic,x86_64,-junix,-smalloc,-sfile,-sdefa
ult,-hcritbit#012varnish-6.6.1 revision e6a8c860944c4f6a7e1af9f40674ea78bbdcdc66#012#012Type 'help' for command list.#012Type 'quit' to close CLI session.
Apr  8 22:42:10 xxx varnishd[10919]: CLI telnet 127.0.0.1 5282 127.0.0.1 6082 Rd ping
Apr  8 22:42:10 xxx varnishd[10919]: CLI telnet 127.0.0.1 5282 127.0.0.1 6082 Wr 200 PONG 1649450530 1.0
Apr  8 22:42:10 xxx varnishd[10919]: CLI telnet 127.0.0.1 5282 127.0.0.1 6082 Rd vcl.load reload_20220408_204210_11818 /etc/varnish/default.vcl
Apr  8 22:42:15 xxx varnishreload[11818]: VCL 'reload_20220408_204210_11818' compiled
Apr  8 22:42:20 xxx varnishreload[11818]: Command: varnishadm -n '' -- vcl.use reload_20220408_204210_11818
Apr  8 22:42:20 xxx varnishreload[11818]: Rejected 400
Apr  8 22:42:20 xxx varnishreload[11818]: CLI communication error (hdr)
Apr  8 22:42:20 xxx systemd[1]: varnish.service: Control process exited, code=exited, status=1/FAILURE
Apr  8 22:42:20 xxx systemd[1]: Reload failed for Varnish Cache, a high-performance HTTP accelerator.

and now ps shows:

vcache     10919  0.0  0.0  19048  5880 ?        SLs  22:38   0:00 /usr/sbin/varnishd -j unix,user=vcache -F -a :80 -T localhost:6082 -f /etc/varnish/default.vcl -p thread_pools=8 -p thread_pool_min=100 -p thread_pool_max=4000 -p workspace_client=128k -p workspace_backend=128k -l 200m -S /etc/varnish/secret -s malloc,256m -s static=file,/data/varnish_storage.bin,80g
vcache     10959  0.4  0.2 84585576 23088 ?      SLl  22:39   0:01 /usr/sbin/varnishd -j unix,user=vcache -F -a :80 -T localhost:6082 -f /etc/varnish/default.vcl -p thread_pools=8 -p thread_pool_min=100 -p thread_pool_max=4000 -p workspace_client=128k -p workspace_backend=128k -l 200m -S /etc/varnish/secret -s malloc,256m -s static=file,/data/varnish_storage.bin,80g

I see process owner was changed to vcache. What is wrong with it? anoder reload will fail too with same reject code.

Answer 1

Can you try removing -j unix,user=vcache from your varnishd runtime command. If I remember correctly, Varnish will automatically drop privileges on the worker process without really needing to explicitly set jailing settings.

If that doesn't work, please also explain which commands you used to start Varnish and reload Varnish.