Auxiliary Copy Backlog Fills a Deduplicated Disk Library
A storage policy writes weekly fulls and daily incrementals to a 60 TB deduplicated disk library and then auxiliary-copies to a 4-drive LTO tape library. Primary retention is 14 days; secondary retention is one year. Deduplication is reported at 5:1. Auxiliary copy is scheduled once daily at 01:00. One tape drive has been offline for a week, library cleaning is overdue, and auxiliary copy jobs have failed for six days with 'media mount timeout.' Twenty-two auxiliary jobs are queued, and the disk library is 94% full. Backups are still succeeding. An administrator proposes deleting the oldest primary jobs to free space immediately and disabling the secondary copy until the tape library is repaired. The RPO for critical applications is 4 hours; the RTO is 8 hours.
1. Why is the disk library full despite 14-day primary retention and 5:1 deduplication? Explain the relationship between primary copy pruning and auxiliary copy.
Read the worked answer
Primary retention is a minimum, not a guarantee that data is deleted at day 14. If the secondary copy has not completed, the primary copy may still be needed as the source for auxiliary copy, so pruning can be held back. That is why a six-day auxiliary backlog can retain data well beyond the expected window. The 5:1 ratio is also an average, not a capacity plan: full backups, changed blocks, metadata, DDB growth, and temporary space for compaction all consume capacity. The administrator's proposal to delete the oldest primary jobs is dangerous because those jobs may not yet be fully copied to tape. Manual deletion can break the secondary copy chain, create gaps in retention, and make restore points unavailable. Disabling the secondary copy does not solve the immediate problem and may violate the one-year retention requirement. The first question is not 'what can we delete?' but 'why has auxiliary copy stopped, and which primary jobs are still eligible to prune?' Check copy precedence and job status before touching data.
2. Evaluate the administrator's proposal. What would you do first, and what temporary measure is safest if the tape library cannot be repaired quickly?
Read the worked answer
Do not delete primary jobs manually and do not disable the secondary copy. First restore tape capability: bring the failed drive back online, complete library cleaning, verify media availability, and confirm the library is not in a failed or reservation state. Then run auxiliary copy in controlled batches, starting with the oldest jobs and critical applications. If the tape library cannot be repaired quickly, a temporary secondary copy to another supported target, such as a separate disk library or approved cloud object storage, can relieve the primary copy dependency, but only if there is enough capacity and the retention design is understood. A selective copy of critical data may protect the RPO while the full backlog is addressed, but it does not replace the missing one-year retention for everything else. Freeing space by expanding the disk library or moving the DDB is safer than deleting primary jobs blindly. Communicate the retention gap explicitly; do not represent the temporary copy as equivalent to the tape copy until it is proven by a restore test.
3. What long-term changes would prevent recurrence? Address scheduling, copy precedence, monitoring, and capacity.
Read the worked answer
Schedule auxiliary copy to run after the backup window and before the next backup cycle, with enough time and drive throughput to clear the expected daily change rate. Set copy precedence so the secondary copy is completed before primary pruning is allowed, and monitor that SLA rather than only monitoring backup success. Add alerting on auxiliary copy queue depth, tape drive state, cleaning overdue, library capacity, and disk library watermarks. Capacity planning should include DDB growth, full backup size, deduplication ratio variance, and the retention of both copies; a 5:1 ratio should not be treated as guaranteed. Add tape drives or a second library if the daily change rate cannot be copied within the available window. Test restores from both the primary and secondary copies, and test a tape library drive failure scenario. Finally, document the acceptable temporary copy target and the approval path for using it, so an operator under pressure does not improvise by deleting primary data.