I am now running yugabyte-2.14.0.0 on 3 VMs, and on each VM there is 1 yb-master and 1 yb-tserver running. Each VM has 8 CPUs and 16 GB RAM. I have just been able to reproduce the issue, but this time for a different table, as seen from the following excerpt from the logs of the PostgreSQL server brought up by the yb-tserver on one of the nodes:
2022-07-20 13:49:27.882 UTC [340693] LOG: statement: CREATE TABLE ptnodedisplaysettings
(
id integer NOT NULL,
color CHARACTER VARYING,
stroke_color CHARACTER VARYING,
stroke_width integer,
size integer,
solid_nofill CHARACTER VARYING,
font_size integer,
label_y_offset integer,
font_weight CHARACTER VARYING,
label_attributes CHARACTER VARYING,
pt_site_name CHARACTER VARYING,
pt_project_name CHARACTER VARYING,
CONSTRAINT pk_pt_node_display_settings PRIMARY KEY (id),
CONSTRAINT pt_node_display_settings_foreign_key_pt_site_constr FOREIGN KEY (pt_site_name, pt_project_name) REFERENCES ptsite(name, pt_project_name) ON DELETE CASCADE
);
I0720 13:49:27.980901 340693 ybccmds.c:483] Creating Table unims.public.ptnodedisplaysettings
2022-07-20 13:49:32.677 UTC [340693] LOG: statement: INSERT INTO ptnodedisplaysettings (id, color, stroke_color, stroke_width, size, solid_nofill, font_size, label_y_offset, font_weight, l
abel_attributes) VALUES (1, ‘#ff8000’, ‘#ff0000’, 2, 6, ‘Solid’, 11, -34, ‘bold’, ‘name’);
2022-07-20 13:49:32.702 UTC [340693] ERROR: could not open relation with OID 17527 at character 13
After this failure, as I expected I was able to describe the table:
mydb=# \d ptnodedisplaysettings
Table "public.ptnodedisplaysettings"
Column | Type | Collation | Nullable | Default
------------------+-------------------+-----------+----------+---------
id | integer | | not null |
color | character varying | | |
stroke_color | character varying | | |
stroke_width | integer | | |
size | integer | | |
solid_nofill | character varying | | |
font_size | integer | | |
label_y_offset | integer | | |
font_weight | character varying | | |
label_attributes | character varying | | |
pt_site_name | character varying | | |
pt_project_name | character varying | | |
Indexes:
"pk_pt_node_display_settings" PRIMARY KEY, lsm (id HASH)
Foreign-key constraints:
"pt_node_display_settings_foreign_key_pt_site_constr" FOREIGN KEY (pt_site_name, pt_project_name) REFERENCES ptsite(name, pt_project_name) ON DELETE CASCADE
In the table UI I saw that the ptnodedisplaysettings table had 1 tablet.
Finally, there were no errors logged for yb-master nor yb-tserver on any of the 3 VMs at the time this happened.
The yb-master was started on each node using a command of the following form:
yb-master --master_addresses :7100,:7100,:7100 --rpc_bind_addresses :7100 --fs_data_dirs /data/vlst/yugabyte/yugabyte-2.14.0.0/data
and the yb-tserver was brought up on each VM with a command of this format:
yb-tserver --tserver_master_addrs :7100,:7100,:7100 \
–rpc_bind_addresses :9100 \
–fs_data_dirs /data/vlst/yugabyte/yugabyte-installed/data \
–start_pgsql_proxy \
–pgsql_proxy_bind_address :5433 \
–ysql_log_statement all \
–ysql_timezone XXXXX \
–pg_yb_session_timeout_ms 900000 \
–cql_proxy_bind_address :9042
Your input is greatly appreciated.
Thank you