View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0018368 | CentOS-8 | kernel | public | 2021-12-02 14:05 | 2022-02-03 22:14 |
Reporter | ksyblast | Assigned To | |||
Priority | normal | Severity | major | Reproducibility | always |
Status | closed | Resolution | won't fix | ||
OS Version | 8.5.2111 | ||||
Summary | 0018368: 4.18.0-348.2.1 memory leak | ||||
Description | We have noticed a significant memory leak using strongswan after upgrading to 4.18.0-348.2.1.el8_5.x86_64 Rollback to 4.18.0-305.3.1 fixes the issue. I am not sure what exactly triggers the leak as I cannot reproduce it on a test server with one ipsec tunnel and low load. | ||||
Steps To Reproduce | 1. strongswan with kernel 4.18.0-348.2.1 (U5.9.4/K4.18.0-348.2.1.el8_5.x86_64) 2. install several (could not reproduce with one) ipsec tunnels + gre interfaces, I can attach configs to test with if needed 3. produce some real load (could not reproduce with simple ping or iperf) 4. memory will slowly leak | ||||
Additional Information | I have tried to investigate the issue with debug kernel and I see kmemleak flooded with the following, that's why I suspect ipsec (attached a bigger sample): unreferenced object 0xffff888034696e80 (size 88): comm "softirq", pid 0, jiffies 4294988744 (age 1475.918s) hex dump (first 32 bytes): 01 00 00 00 01 00 00 00 00 00 00 00 4e 02 a0 9c ............N... 80 ba ad 2f 80 88 ff ff cb 93 4b 97 ff ff ff ff .../......K..... backtrace: [<000000002f241681>] kmem_cache_alloc+0xef/0x300 [<0000000033b21ffc>] secpath_dup+0x23/0x1d0 [<0000000004ded9bb>] secpath_set+0x9f/0x160 [<000000001476d24b>] xfrm_input+0x29c/0x2850 [<00000000bd5f25fe>] xfrm4_esp_rcv+0x9f/0x190 [<000000001afc6bf7>] ip_protocol_deliver_rcu+0x5ae/0x7d0 [<00000000913ba334>] ip_local_deliver_finish+0x222/0x330 [<000000005bd92736>] ip_local_deliver+0x326/0x410 [<00000000e031e119>] ip_rcv+0xa7d/0x123d [<00000000f9ed8d16>] __netif_receive_skb_core+0x2051/0x3330 [<00000000b24f5f1c>] netif_receive_skb_internal+0xed/0x340 [<000000001ea19f44>] napi_gro_receive+0x27f/0x3c0 [<00000000d2e61b66>] xennet_poll+0x1a1b/0x2ac0 [xen_netfront] [<00000000cccfb516>] __napi_poll+0x9f/0x520 [<0000000073da1556>] net_rx_action+0x36e/0x980 [<00000000f705dcfe>] __do_softirq+0x1cb/0xa4f | ||||
Tags | No tags attached. | ||||
You need to report the bug (if such it is) on bugzilla.redhat.com. CentOS only rebuilds what is put out by RH for RHEL and does not fix bugs itself so it is pointless reporting them here. | |
Besides, CentOS 8 ended its life on December 31, 2021 and, therefore, is no longer supported. | |
Date Modified | Username | Field | Change |
---|---|---|---|
2021-12-02 14:05 | ksyblast | New Issue | |
2021-12-02 14:05 | ksyblast | File Added: kmemleak.tar.gz | |
2022-01-01 16:25 | TrevorH | Note Added: 0038778 | |
2022-01-08 18:51 | toracat | Note Added: 0038780 | |
2022-02-03 22:14 | toracat | Status | new => closed |
2022-02-03 22:14 | toracat | Resolution | open => won't fix |
2022-02-03 22:14 | toracat | Description Updated | |
2022-02-03 22:14 | toracat | Steps to Reproduce Updated | |
2022-02-03 22:14 | toracat | Additional Information Updated |