# Segfaults on old ScyllaDB cluster, how can I upgrade it?

**URL:** <https://forum.scylladb.com/t/segfaults-on-old-scylladb-cluster-how-can-i-upgrade-it/5159>\
**Category:** ScyllaDB\
**Tags:** troubleshooting, administration, backup-restore, upgrade\
**Created:** [October 28, 2025, 9:15am UTC](https://forum.scylladb.com/t/segfaults-on-old-scylladb-cluster-how-can-i-upgrade-it/5159 "2025-10-28T09:15:10Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Guy](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.scylladb.com/guy/32/9_2.png) [@Guy](https://forum.scylladb.com/u/Guy)\
**Post date:** [October 28, 2025, 9:15am UTC](https://forum.scylladb.com/t/segfaults-on-old-scylladb-cluster-how-can-i-upgrade-it/5159/1 "2025-10-28T09:15:11Z")

</div>

> [@](#):
>
> **[Originally from the User Slack](https://scylladb-users.slack.com/)**
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @SabinSabin\*\*:\*\* I have really old Scylla Server (`Scylla version 4.2.1-0.20201108.4fb8ebccff`) which is “running” at work. Recently, I have encountered issues where there’s random segfaults on shard X. How can I debug and possibly fix it? I want to migrate it to newer version but this issue is holding me back  
> I am mostly seeing these on the `journalctl` for Scylla Server
> 
> ```auto
> Oct 22 04:53:29 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: [shard 0] compaction - Compacting [/var/lib/scylla/data/system/clients-ca0f635d863036098d93a1fc06f2a5e5/mc-12922-big-Data.db:level=0, /var/lib/scylla/data/system/clients-ca0f635d863036098d93a1fc06f2a5e5/mc-12936-big-Data.db:level=0,]
> Oct 22 04:53:30 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: [shard 0] compaction - Compacted 2 sstables to [/var/lib/scylla/data/system/clients-ca0f635d863036098d93a1fc06f2a5e5/mc-12950-big-Data.db:level=0,]. 93kB to 17kB (~18% of original)
> in 118ms = 144kB/s. ~256 total partitions merged to 1.
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: Segmentation fault on shard 0.
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: Backtrace:
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002ef3122
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e976a0
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e97945
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e979e0
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x00007f548046ca8f
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000f2b223
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000f2b61f
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001087aea
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000fe2e23
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001000ab5
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001005c1b
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001007949
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x000000000100807b
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001007a5c
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x000000000116f281
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x000000000117171d
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001171c9a
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x00000000010d0f97
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001089760
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000001092a23
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x00000000010c2f22
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000e380a9
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000e3ac0c
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e2b3e6
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e94a47
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e94dbe
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002ecbe9d
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e28c7a
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000002e2932e
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000da8b40
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: /opt/scylladb/libreloc/libc.so.6+0x0000000000027041
> Oct 22 04:53:39 <http://scylladb-1.mycompany.com|scylladb-1.mycompany.com> scylla[1841]: 0x0000000000cc4a2d
> 
> ```
> 
> This crashes the service and the kernel on… Ubuntu 18.04 TL@avi
> 
> ![avi](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/1X/b5e08becda8d9df1abff1d3a2d9ffe1126ac3fc6.jpeg) @avi\*\*:\*\* It is highly recommended to upgrade both the server OS and the ScyllaDB soft@Sabinare
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @Sabin\*\*:\*\* If I create a new instance with recent version of ScyllaDB, would it be able to join the@avicluster?
> 
> ![avi](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/1X/b5e08becda8d9df1abff1d3a2d9ffe1126ac3fc6.jpeg) @avi\*\*:\*\* It’s not recommended, especially for such old versions  
> Upgrade in place, one version at a time, until you reach a supported version, and don’t let it lag i@Sabin the future
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @Sabin\*\*:\*\* I am worried about the data. Also, I don’t have much experience wi@avih the DB itself.
> 
> ![avi](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/1X/b5e08becda8d9df1abff1d3a2d9ffe1126ac3fc6.jpeg) @avi\*\*:\*\* Take a backup.
> 
> Leaving it like that will cause it to rot until one day yo@Sabin cannot recover it.
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @Sabin\*\*:\*\* Is there a way to fix this error apart from upgrading? I have backup but recoveri@avig it takes way too long.
> 
> ![avi](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/1X/b5e08becda8d9df1abff1d3a2d9ffe1126ac3fc6.jpeg) @avi\*\*:\*\* Try to decode it on @Sabintp://backtrace.scylladb.com
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @Sabin\*\*:\*\* I wonder if I can do rolling up@avirade of ScyllaDB from 4.x to 6.x
> 
> ![avi](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/1X/b5e08becda8d9df1abff1d3a2d9ffe1126ac3fc6.jpeg) @avi\*\*:\*\* No, we only test one minor version at a time
> 
> ![Sabin](https://us1.discourse-cdn.com/flex016/uploads/scylladb/original/2X/1/1b2b271dfb28791c3a9f779604b13028ba37027d.jpeg) @Sabin\*\*:\*\* That will take time. The node is now up as the actual issue was with hardware (faulty RAM sticks). I will try to get the data off the nodes and try to restore it on newer version on another cluster.