Loading...
Loading...
The memory in the device remains increasing until start to report high memory utilization error logs, the high memory utilization may be detected by a monitoring tool. From the memory monitoring (in this case the device was reloaded before to monitor the memory), most of the memory utilization increment is due to vman process: Comparison after 15 days: ++++++++ #show platform software status control-processor brief ++++++++ The column used for the comparison is 'Used' The output is in KB SLOT TOTAL BEFORE AFTER CHANGE RP0 15823064 4322164 4375292 53128 ++++++++ #show processes memory platform sorted ++++++++ The column used for the comparison is 'RSS' The output is in KB NAME BEFORE AFTER INCREMENT vman 55096 77668 22572 <<< cpp_cp_svr 240728 248332 7604 fman_fp_image 220200 225028 4828 cman_fp 47216 51120 3904 fman_rp 145152 148520 3368 Vman process is allocating memory into a smap called heap, the memory allocated is increasing constantly and at certain point most of the memory on vman is consumed by this smap, (there are some smap called heap but the increment is only seen on one): show processes memory platform detailed name vman smaps location r0 60bd17f4b000-60bd1a834000 rw-p 00000000 00:00 0 [heap] Size: 41892 kB KernelPageSize: 4 kB MMUPageSize: 4 kB Rss: 41796 kB <<< #show process memory platform sorted Pid Text Data Stack Dynamic RSS Name ---------------------------------------------------------------------- 3409 402216 1174084 136 456 1174084 linux_iosd-imag [skip] 12546 893 77668 136 41892 77668 vman From the outputs above, The smap heap is consuming 41 MB while vman process is consuming 77 MB.
This issue was seen on a C8500 running in 17.12.2, therefore the issue may affect other platforms running IOS-XE 17.12.2 or 17.12.1 The issue has been reproduced in C8300 and C8500 running in 17.12.2, therefore we are suspecting all the C8K platforms are affected. The issue starts to happen after a the device gets connected to the Thousand eyes cloud, the issue happens when the connection is done directly or through a proxy
Reloading the device will release the memory consumed by the leak but the memory will start to increase again.
The issue was detected on a device running in autonomous mode, we are checking if this may affect controller mode as well.
Click on a version to see all relevant bugs
Cisco Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.