On Friday, October 27, 2017 at 10:07:32 AM UTC+8, Tam Angustia wrote: > Hi all, > > Ive noticed an issue regarding a client that is going on a backup operation. > First of all, Ive been searching why the storage daemon issues a broken pipe > when I am only backing up between VLANs. > Ive narrowed down the problem to a network issue. > Next Ive compared clients who terminated successfully and saw their MTU was > set to 1500 as default. Some clients need a 9000 mtu for other purpose as > well. > When I edited the mtu on the client to revert back to 1500, it now continues > to backup the said client without further issues. > > First thing is our setup > I am using an external SAS connection (DELL MD3200 storage stack) to my > bareos server with also having an mtu of 9000 (jumbo frames) > > I am also using an external percona cluster database for the catalogs > > I also disabled batch inserts as ive encountered errors when Ive set the > database of bareos to the external percona cluster. > > Can anyone confirm if the bareos has problems regarding MTUs?
Anyone? Ill be tackling the issue on the maximum network buffer size. As of now im still verifying the relationship between the MTU and the maximum TCP window. But Im pretty sure the MTU has something to do with the backup jobs that hang. Anyone from the bareos team that had replicated this? -- You received this message because you are subscribed to the Google Groups "bareos-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/d/optout.
