Re: daos_io_server fails if daos_server.log is missing
|
Hello Kevan,
Can you supply your server configuration file please, the ownership changes that are being triggered are part of a previous attempt to work around requiring elevated permissions when formatting storage prior to wanting to run daos_io_server as a normal user.
This code will be removed in a current WIP PR of mine as this requirement will not exist when the formatting operations are performed by the standalone privileged binary (daos_admin). The requisite work will be completed soon.
I am imagining you have “username” set in your config file, try removing that (or groupname) if set.
Apologies for any inconvenience.
Regards, Tom Nabarro – DCG/ESAD M: +44 (0)7786 260986 Skype: tom.nabarro
From: daos@daos.groups.io <daos@daos.groups.io>
On Behalf Of Kevan Rehm
Sent: Friday, December 13, 2019 8:40 PM To: daos@daos.groups.io Subject: [daos] daos_io_server fails if daos_server.log is missing
Greetings,
I’d like to mention a bug that I tripped on today, in case I’m the first to notice it. As background, I followed the instructions from about a week ago on how to chown/chgrp/chmod the binaries daos_admin and daos_server to make them setuid so that daos_server could run as user daos. If I do try to run daos_server as user daos with a “file” nvme device, it fails in spdk initialization, I’m not sure why yet, but…
If I instead start daos_server as root, it dies if file /tmp/daos_server.log does not exist.
changing file ownership hl-c507 ERROR 2019/12/13 13:20:14 changing file ownership: recursive chown /tmp/daos_server_svr1.log: accessing path /tmp/daos_server_svr1.log: lstat /tmp/daos_server_svr1.log: no such file or directory
Routine chownAll() checks the error code from a failed chown for !os.IsNotExist(err) but routine ChownR is apparently returning some other error, so the test fails, and the daos_io_server exits.
I will try touching the file on each server to get around this. If there’s not a Jira already, let me know.
Thanks, Kevan
P.S. here is the daos_control.log:
DEBUG 13:20:13.942072 start.go:145: Switching control log level to DEBUG DEBUG 13:20:13.942860 server.go:63: cfg: &server.Configuration{ControlPort:10000, TransportConfig:(*security.TransportConfig)(0xc00018d420), Servers:[]*ioserver.Config{(*ioserver.Config)(0xc000001b00)}, BdevInclude:[]string(nil), BdevExclude:[]string(nil), NrHugepages:4096, ControlLogMask:3, ControlLogFile:"/tmp/daos_control.log", ControlLogJSON:false, HelperLogFile:"", UserName:"daos", GroupName:"daos_grp", RecreateSuperblocks:true, SystemName:"daos_krehm", SocketDir:"/var/run/daos_server", Fabric:ioserver.FabricConfig{Provider:"ofi+verbs;ofi_rxm", Interface:"", InterfacePort:0, PinnedNumaNode:(*uint)(nil)}, Modules:"", AccessPoints:[]string{"hl-c500:10000"}, FaultPath:"", FaultCb:"", Hyperthreads:false, Path:"/home/users/daos/daos/install/etc/daos_server.yml", ext:(*server.ext)(0xc00012e880), NvmeShmID:227167734, validateProviderFn:(server.networkProviderValidation)(0x5e1c00), validateNUMAFn:(server.networkNUMAValidation)(0x5e3110)} DEBUG 13:20:13.942904 netdetect.go:749: Calling ValidateProviderConfig with ib0, ofi+verbs;ofi_rxm DEBUG 13:20:13.942945 netdetect.go:793: Input provider string: ofi+verbs;ofi_rxm DEBUG 13:20:14.077492 netdetect.go:825: There are 0 hfi1 devices in the system DEBUG 13:20:14.077543 netdetect.go:764: Device ib0 supports provider: ofi+verbs;ofi_rxm DEBUG 13:20:14.077769 netdetect.go:901: Validate network config -- given numaNode: 0 DEBUG 13:20:14.208993 netdetect.go:926: The NUMA node for device ib0 matches the provided value 0. Network configuration is valid. DEBUG 13:20:14.209736 config.go:410: Active config saved to /home/users/daos/daos/install/etc/.daos_server.active.yml (read-only) hl-c507 INFO 2019/12/13 13:20:14 DAOS control server (pid 24144) listening on 0.0.0.0:10000 hl-c507 INFO 2019/12/13 13:20:14 Waiting for I/O server instance storage to be ready... DEBUG 13:20:14.309826 instance.go:177: /mnt/daos/svr1: checking formatting DEBUG 13:20:14.319004 instance.go:195: /mnt/daos/svr1 (ram) needs format: false DEBUG 13:20:14.319035 superblock.go:107: /mnt/daos/svr1: checking superblock DEBUG 13:20:14.319052 instance.go:177: /mnt/daos/svr1: checking formatting DEBUG 13:20:14.327795 instance.go:195: /mnt/daos/svr1 (ram) needs format: false DEBUG 13:20:14.329552 drpc_server.go:85: Quitting listener hl-c507 ERROR 2019/12/13 13:20:14 /var/run/daos_server/daos_server.sock: failed to accept connection: accept unixpacket /var/run/daos_server/daos_server.sock: use of closed network connection DEBUG 13:20:14.329973 main.go:68: lstat /tmp/daos_server_svr1.log: no such file or directory accessing path /tmp/daos_server_svr1.log github.com/daos-stack/daos/src/control/server.(*ext).chownR.func1 /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/external.go:101 path/filepath.Walk /usr/lib/golang/src/path/filepath/path.go:402 github.com/daos-stack/daos/src/control/server.(*ext).chownR /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/external.go:99 github.com/daos-stack/daos/src/control/server.chownAll /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/ownership.go:95 github.com/daos-stack/daos/src/control/server.changeFileOwnership /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/ownership.go:122 github.com/daos-stack/daos/src/control/server.Start /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/server.go:207 main.(*startCmd).Execute /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/start.go:165 main.parseOpts.func1 /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:107 github.com/daos-stack/daos/src/control/vendor/github.com/jessevdk/go-flags.(*Parser).ParseArgs /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/vendor/github.com/jessevdk/go-flags/parser.go:314 main.parseOpts /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:115 main.main /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:127 runtime.main /usr/lib/golang/src/runtime/proc.go:203 runtime.goexit /usr/lib/golang/src/runtime/asm_amd64.s:1357 recursive chown /tmp/daos_server_svr1.log github.com/daos-stack/daos/src/control/server.chownAll /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/ownership.go:97 github.com/daos-stack/daos/src/control/server.changeFileOwnership /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/ownership.go:122 github.com/daos-stack/daos/src/control/server.Start /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/server/server.go:207 main.(*startCmd).Execute /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/start.go:165 main.parseOpts.func1 /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:107 github.com/daos-stack/daos/src/control/vendor/github.com/jessevdk/go-flags.(*Parser).ParseArgs /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/vendor/github.com/jessevdk/go-flags/parser.go:314 main.parseOpts /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:115 main.main /home/users/daos/daos/build/src/control/src/github.com/daos-stack/daos/src/control/cmd/daos_server/main.go:127 runtime.main /usr/lib/golang/src/runtime/proc.go:203 runtime.goexit /usr/lib/golang/src/runtime/asm_amd64.s:1357 changing file ownership hl-c507 ERROR 2019/12/13 13:20:14 changing file ownership: recursive chown /tmp/daos_server_svr1.log: accessing path /tmp/daos_server_svr1.log: lstat /tmp/daos_server_svr1.log: no such file or directory
--------------------------------------------------------------------- This e-mail and any attachments may contain confidential material for
|
|
|