setup_jobserver fails with 'Length of node exceeds maximum length 32'

[wim.ve...@billinghouse.nl]

Hi,

I have installed schedulix on an AWS system. Now I try to actually get a jobserver running, but it fails due to a long name.

I tried with the following cmd line:

bash-4.2$ ./setup_jobserver -n LHSERVER -w xxx -o PUBLIC -H localhost -P 2506 -h 45500 -W /opt/schedulix/workdirs/localhost -c -u SYSTEM -p xxx

I get the following output:

The full qualified name of the jobserver will be GLOBAL.'SCHEDULIX-EU-WEST-1A-1'.'SCHEDULIX'.'LHSERVER'

Is this correct? [Y/n] y

The HTTPPORT that will be used for logfile display is 45500

This can be changed at any time in the config tab of the jobserver

Connect

CONNECT_TIME : 20 Sep 2018 13:38:54 GMT

Connected

[SYSTEM@localhost:2506] SDMS> 

End of Output

[SYSTEM@localhost:2506] SDMS> ;

[SYSTEM@localhost:2506] SDMS> begin multicommand

create or alter scope GLOBAL.'SCHEDULIX-EU-WEST-1A-1'

with

group = 'ADMIN',

config = (

'REPOHOST' = 'localhost',

'REPOPORT' = '2506',

'BOOTTIME' = 'NONE',

'USEPATH' = 'true',

'JOBEXECUTOR' = '/opt/schedulix/schedulix/bin/jobexecutor',

'DEFAULTWORKDIR' = '/opt/schedulix/workdirs/localhost',

'VERBOSELOGS' = 'true',

'HTTPHOST' = 'schedulix-eu-west-1a-1.dev09.nova.essent.be',

'NAME_PATTERN_LOGFILES' = '.*\.log',

'ENV' = (

'SDMSHOST' = 'SDMSHOST',

'JOBID' = 'JOBID',

'SDMSPORT' = 'SDMSPORT',

'KEY' = 'KEY'

)

);

create or alter scope GLOBAL.'SCHEDULIX-EU-WEST-1A-1'.'SCHEDULIX'

with

group = 'PUBLIC';

create or alter job server GLOBAL.'SCHEDULIX-EU-WEST-1A-1'.'SCHEDULIX'.'LHSERVER'

with

group = 'PUBLIC',

password = 'xxx',

node = 'schedulix-eu-west-1a-1.dev09.nova.essent.be',

config = (

'JOBFILEPREFIX' = '/opt/schedulix/taskfiles/SCHEDULIX-EU-WEST-1A-1/schedulix/LHSERVER/LHSERVER-'

, HTTPPORT = '45500'

);

create or alter named resource RESOURCE.'STATIC'

with

group = 'PUBLIC',

usage = CATEGORY;

create or alter named resource RESOURCE.'STATIC'.'HOSTS'

with

group = 'PUBLIC',

usage = CATEGORY;

create or alter named resource RESOURCE.'STATIC'.'HOSTS'.'SCHEDULIX-EU-WEST-1A-1'

with

group = 'PUBLIC',

usage = STATIC;

create or alter named resource RESOURCE.'STATIC'.'USERS'

with

group = 'PUBLIC',

usage = CATEGORY;

create or alter named resource RESOURCE.'STATIC'.'USERS'.'SCHEDULIX'

with

group = 'PUBLIC',

usage = STATIC;

create or alter named resource RESOURCE.'STATIC'.'SERVERS'

with

group = 'PUBLIC',

usage = CATEGORY;

create or alter named resource RESOURCE.'STATIC'.'SERVERS'.'LHSERVER'

with

group = 'PUBLIC',

usage = STATIC;

create or alter resource RESOURCE.'STATIC'.'HOSTS'.'SCHEDULIX-EU-WEST-1A-1' in GLOBAL.'SCHEDULIX-EU-WEST-1A-1'

with

group = 'PUBLIC',

online;

create or alter resource RESOURCE.'STATIC'.'USERS'.'SCHEDULIX' in GLOBAL.'SCHEDULIX-EU-WEST-1A-1'.'SCHEDULIX'

with

group = 'PUBLIC',

online;

create or alter resource RESOURCE.'STATIC'.'SERVERS'.'LHSERVER' in GLOBAL.'SCHEDULIX-EU-WEST-1A-1'.'SCHEDULIX'.'LHSERVER'

with

group = 'PUBLIC',

online;

create or alter environment 'SCHEDULIX_LHSERVER@SCHEDULIX-EU-WEST-1A-1'

with

resources = (

RESOURCE.'STATIC'.'HOSTS'.'SCHEDULIX-EU-WEST-1A-1',

RESOURCE.'STATIC'.'USERS'.'SCHEDULIX',

RESOURCE.'STATIC'.'SERVERS'.'LHSERVER'

);

grant USE, VIEW

on environment 'SCHEDULIX_LHSERVER@SCHEDULIX-EU-WEST-1A-1'

to 'PUBLIC';

end multicommand;

ERROR:01112141528, Error in Statement 3 (CreateScope) : (Scope) Length of node exceeds maximum length 32

[SYSTEM@localhost:2506] SDMS> 

ERROR: something went wrong while trying to create the jobserver on the server side

the created files and directories will be removed. After eliminating the error cause

the script can be executed again.

The limit of 32 seems exceptionally small. The database definition seems to allow 64 length:

CREATE TABLE scope (

    id numeric(20,0) NOT NULL,

    name character varying(64) NOT NULL,

    owner_id numeric(20,0) NOT NULL,

    parent_id numeric(20,0),

    type integer NOT NULL,

    is_terminate integer,

    has_alteredconfig integer,

    is_suspended integer,

    is_enabled integer,

    is_registered integer,

    state integer,

    passwd character varying(64),

    salt character varying(64),

    method integer NOT NULL,

    pid character varying(32),

    node character varying(32),

    errmsg character varying(256),

    last_active numeric(20,0),

    creator_u_id numeric(20,0) NOT NULL,

    create_ts numeric(20,0) NOT NULL,

    changer_u_id numeric(20,0) NOT NULL,

    change_ts numeric(20,0) NOT NULL,

    inherit_privs numeric(20,0) NOT NULL

);

I use version 2.8.

I will try to do it manually, but it is PITA

Regards,

Wim Veldhuis.

[Dieter Stubler]

Hi Wim,

the problem is not the name of the jobserver but the node name which is indeed defined to be 32 Chars only.
We didn' had a host name longer than 32 chars yet.

...

node = 'schedulix-eu-west-1a-1.dev09.nova.essent.be',

...

We'll check whether we can change this to be 64 chars in future releases.

Regards
Dieter

[wim.ve...@billinghouse.nl]

Thanks for the feedback.

Now I can work around the issue most likely. I think we can use a shorter name with an alias.

The good news is we are, after some testing in may, finally progressing with setting up a system for our environment. There are still some things to learn, but I feel we are not wanting to do the impossible.

Biggest issue at the moment is infrastructure (proxies, etc). I might have questions about that later on, but I will start a new thread then.