☰

libnss-pgsql2 - getent 组失败

tag-icon • tag-icon linux login postgresql nss getent
libnss-pgsql2 - getent 组失败

我已经在 Linux 服务器（ubuntu 16.04）上设置了 pam 和 nss，以从远程 postgresql 数据库获取用户和组。在服务器上登录没问题，但无法从数据库中加载组。

如果表 group_table 中存在条目，getent group 将失败并显示以下消息'第 3 列超出范围 0..2'。 nss-pgsql.conf 中使用的所有 sql 似乎都很好。

我使用以下数据库表：

表密码

用户组

组表

nss-pgsql.conf 的内容：

connectionstring        = hostaddr=123.456.789.10 dbname=nobodyexpects user=the password=spanishinquisition connect_timeout=1
getgroupmembersbygid    = SELECT username FROM passwd_table WHERE gid = $1
getpwnam        = SELECT username, passwd, gecos, homedir, shell, uid, gid FROM passwd_table WHERE username = $1
getpwuid        = SELECT username, passwd, gecos, homedir, shell, uid, gid FROM passwd_table WHERE uid = $1
allusers        = SELECT username, passwd, gecos, homedir, shell, uid, gid FROM passwd_table
getgrnam        = SELECT groupname, passwd, gid FROM group_table WHERE groupname = $1
getgrgid        = SELECT groupname, passwd, gid FROM group_table WHERE gid = $1
groups_dyn      = SELECT ug.gid FROM passwd_table JOIN usergroups ug USING (uid) where username = $1 and ug.gid <> $2
allgroups       = SELECT groupname, passwd, gid  FROM group_table

如果表 group_table 不包含任何数据，我会收到一条消息，指出找不到组的名称，并且 getent 组与本地组一起使用。希望有人能帮忙。

答案1

发现 group-SQL 需要额外的用户名列表。将 SQL 更改为

getgrnam        = SELECT g.groupname, 'x' AS passwd, g.gid, ARRAY(SELECT p.username FROM passwd_table p INNER JOIN usergroups ug ON ug.uid=p.uid WHERE ug.gid = g.gid) AS members FROM group_table g WHERE g.groupname = $1
getgrgid        = SELECT g.groupname, 'x' AS passwd, g.gid, ARRAY(SELECT p.username FROM passwd_table p INNER JOIN usergroups ug ON ug.uid=p.uid WHERE ug.gid = g.gid) AS members FROM group_table g WHERE g.gid = $1
groups_dyn      = SELECT ug.gid FROM passwd_table JOIN usergroups ug USING (uid) where username = $1 and ug.gid <> $2
allgroups       = SELECT groupname, passwd, gid, ARRAY(SELECT p.username FROM passwd_table p INNER JOIN usergroups ug ON ug.uid=p.uid WHERE ug.gid = g.gid) AS members  FROM group_table

解决问题。

相关内容