IBM S2000-025 IBM AIX v7.3 Administrator Specialty Exam Practice Test

AIX 7.2 and earlier shipped both NTP version 3 and NTP version 4 binaries, selected through symbolic links in /usr/sbin. Starting with AIX 7.3, NTP version 3 is removed. To stay compatible with earlier releases, the familiar command names remain as symbolic links that point to the NTP version 4 binaries, so the same commands continue to work.

Technical Deep Dive: For example, /usr/sbin/xntpd is a link to /usr/sbin/ntp4/ntpd4; confirm this with ls -ld /usr/sbin/xntpd and xntpd --version. The daemon is still managed through the SRC with startsrc -s xntpd and stopsrc -s xntpd, and status is checked with lssrc -ls xntpd and ntpq -p. Moving to version 4 also brings stronger key authentication (SHA1 and SHA256 in addition to MD5). Existing /etc/ntp.conf files should be reviewed after migration, but the commands themselves are not broken, which rules out option D. NTP version 5 is not a supported protocol level on AIX.

In the extended drive report produced by iostat -D, tps is the number of transfers (I/O requests) per second issued to the physical disk or tape, and avgtime, in the queue section, is the average time a transfer request spends waiting in the wait queue.

Technical Deep Dive: The other descriptions are incorrect. sqfull is the number of times per second that the service queue became full, meaning the disk is not accepting more requests, not a count over the last hour. avgserv (in the read and write sections) is the average service time per transfer, not a queue length. avgwqsz is the average wait-queue size expressed as a number of requests, not bytes, and avgsqsz is the matching figure for the service queue. A non-zero sqfull together with a growing avgtime points to a queue_depth limit on the hdisk or a saturated storage path; review it with lsattr -El hdiskN -a queue_depth.

The umask value specifies the permission bits that are removed from newly created files and directories. A umask of 077 removes all permissions for group and others, so new files are created with mode 600 (rw-------) and new directories with mode 700 (rwx------), accessible only by their owner.

Technical Deep Dive: Files start from a base mode of 666 and directories from 777 before the umask is applied. With 022 (the common default) group and others can still read; with 027 the group retains read access to files and directories; and with 047 the group still receives write access to new files (mode 620), so none of these three values keeps data private to the owner. To make the setting persistent for all users, set umask in /etc/security/user (the umask attribute under the default stanza, using chsec -f /etc/security/user -s default -a umask=077) or in /etc/profile; for one user, use chuser umask=077 < user > .

The mkitab command adds a record to /etc/inittab. The record uses the format Identifier:RunLevel:Action:Command, so myservice:2:once:/usr/bin/myservice defines a record for run level 2 whose command is started once when init enters that run level.

Technical Deep Dive: Use mkitab -i < existing_id > "..." to insert the new record after a specific entry, lsitab -a to list records, chitab to change an existing record, and rmitab myservice to remove it. Never edit the file with a redirect: the first option would overwrite the whole inittab with a single line and prevent the system from booting normally. rcupdate is not an AIX command, and chitab modifies an existing record rather than adding one. The action field can be once, wait, respawn, boot, bootwait and others; respawn restarts the command if it exits. For application services, an SRC subsystem or a start script under /etc/rc.d is often preferred over inittab.

AIX exposes its system-wide operating system attributes (the sys0 device) through the ODM, and lsattr -El sys0 displays their current values, for example maxuproc, ncargs and iostat. It is the general-purpose command in this list for showing system tunable values, and lsattr -El < device > likewise shows tunables of adapters and disks, such as queue_depth.

Technical Deep Dive: AIX also has dedicated tuning commands for particular resource groups: vmo (virtual memory), ioo (I/O), no (network), nfso (NFS), schedo (CPU scheduler) and raso (RAS); each displays its own tunables with -a, or with -L for a table of current, default, boot and range values. These do not cover every system tunable, and vmo in particular lists only VMM parameters. The tunsave, tunrestore and tuncheck commands manage tunable stanza files. sysctl is a Linux and BSD kernel-parameter tool that is not part of AIX, and lsdev lists devices rather than their tunable values.

The svmon command captures a snapshot of virtual-memory and real-memory use. It can report system-wide, per-process, per-segment and per-command detail, and with the -i flag it repeats the report at a chosen interval and count, so memory behaviour can be observed over a period of time.

Technical Deep Dive: Examples are svmon -G -i 5 12 (global figures every 5 seconds, 12 times), svmon -P < pid > (one process), svmon -S (segments) and svmon -C < command > . The report separates working (computational) pages, persistent file pages and client file pages, and shows pinned, in-use and paging-space figures. iostat reports CPU and disk I/O activity, lsmem shows how much memory is configured and its logical memory blocks (not usage), and memstat is not part of the AIX base performance toolset. vmstat 5 is a lighter alternative for continuous trends.

AIX tar archives members under the path name used when the archive was created. When absolute path names were archived, a plain extraction writes back to the original location and would replace the live configuration file. The pax command can rewrite member names during extraction with the -s substitution option; the expression /^\//.\// replaces the leading slash with ./, so the file is extracted relative to the current directory (/tmp/newfiles) instead of over /app_dir/appl/config/master.cfg.

Technical Deep Dive: After extraction the administrator can compare the restored copy with the live file (diff) and merge only the required changes. Option A uses tar -t, which only lists archive contents and restores nothing. Option B is not valid AIX tar syntax, because the -o flag is not a destination-directory option. Option D is wrong because the restore command reads archives created by the backup command (by name or by inode) and cannot read tar archives. Always check archive paths first with tar -tvf or pax -f < archive > before restoring.

AIX records console output in a circular log that uses the alog format, and the alog command is the standard tool for reading it. Running alog -t console -o prints the console log, and alog -t boot -o prints the boot log; alog -L lists every log type defined in the alog configuration database.

Technical Deep Dive: The console log (by default /var/adm/ras/conslog) wraps when it reaches its maximum size, so the oldest entries are overwritten first. Its size and verbosity can be changed persistently with alog -C -t console. dmesg reads a kernel ring buffer on Linux and BSD systems and is not part of the AIX base operating system. syslog is the logging facility and daemon (syslogd), not a viewer for console text, and errpt reports entries from the AIX error log maintained by errdemon, which is a different data source.

Each entry in /etc/hosts uses the format Address HostName [Alias ...], one record per line. Appending the line 192.168.1.10 appserver with the > > redirection operator adds the mapping without touching the existing entries.

Technical Deep Dive: Use > > (append), never > , which would overwrite the whole file. AIX also provides the hostent command, whose syntax is hostent -a 192.168.1.10 -h appserver, and smitty mkhostent for the menu approach; the form shown in option D has the wrong flags. addhost and edithosts are not AIX commands. If lookups should prefer the local file, check /etc/netsvc.conf (for example hosts=local,bind4). Verify the entry with host appserver or ping appserver.

The cron daemon runs recurring jobs defined in crontab files. The five time fields are minute, hour, day of month, month and day of week, so 0 5 * * 1 /usr/local/bin/report.sh runs the script at 05:00 every Monday (day of week 0 is Sunday, 1 is Monday).

Technical Deep Dive: crontab -e edits the invoking user's table, crontab -l lists it, and access is governed by /var/adm/cron/cron.allow and cron.deny. Cron jobs run with a minimal environment, so scripts should use absolute paths and set any required variables; output not redirected is mailed to the owner. The at command runs a job once, and batch runs a job once when system load permits, so neither provides a weekly schedule. Scripts in /etc/rc.d are executed at system startup and shutdown, not on a calendar schedule.

The odmget command retrieves objects from one or more ODM object classes and accepts a selection criterion, so it can return the stored records for one specific device. Customized device data is kept in classes such as CuDv (customized devices) and CuAt (customized attributes).

Technical Deep Dive: Examples are odmget -q "name=hdisk0" CuDv for the device record and odmget -q "name=hdisk0" CuAt for its attribute values; the predefined (default) data is in PdDv and PdAt. The ODM files are stored in /etc/objrepos and are read and modified with the odmget, odmadd, odmchange and odmdelete commands; direct edits are discouraged. lsattr and lsdev are high-level tools that read the same data but present formatted device attributes rather than the raw ODM objects, and odmshow displays the class definition (its descriptors), not the stored objects. lsmget is not an AIX command.

The savevg -T flag creates the backup by using snapshots, which applies to JFS2 file systems only. External JFS2 snapshots are created for the file systems in the volume group, the data is backed up from the snapshot images, and the snapshots are removed when the backup completes.

Technical Deep Dive: A snapshot is a point-in-time image of a JFS2 file system, so the backup does not require the system or application to be taken offline. The snapshot logical volume typically needs about 2 to 15 percent of the file system size, so the volume group must have enough free physical partitions. The volume group must be varied on and its file systems mounted. For database workloads, application-level preparation (such as a database hot-backup or write-suspend mode) is still good practice to guarantee application consistency, but it does not require unmounting. The other choices cause downtime (unmount, reboot) or are not valid procedures.

The mkfs command builds a new file system on a specified logical volume or device, so it is the command from this list that creates the file system. The backing logical volume belongs to datavg and must exist before the file system is made on it.

Technical Deep Dive: In everyday administration the crfs command is the preferred one-step method because it also creates the logical volume, builds the file system and adds the /etc/filesystems stanza, for example crfs -v jfs2 -g datavg -m /testfs -a size=1G -A yes, followed by mount /testfs (the SMIT path is smitty crjfs2). chfs only changes attributes of an existing file system, such as its size with chfs -a size=+1G /testfs. addfs is not an AIX command, and mke2fs builds Linux ext2/ext3 file systems that AIX does not use.