Hacker News new | ask | show | jobs
by mesrik 15 days ago
Sure I remember, but since I purchased Spinrite doing lowlevel was needed just once while changing ST-506 controller to a different type or for a new disk, former was quite much rarer needed.

Then even after std lowlevel it was worth using Spinrite to check if interleave value was proper. And if it wasn't it was worth letting it first before anything else. Same when changing a faster CPU as it could speed up IO so much that no interleave would not needed any more and get faster IO.

Spinrite was such a great tool and time saver fixing or making preventive periodic maintenance to customers disks, even though it chugged hours even 30M disks. And just because not to take absolutely any risk it was necessary to make full backup first, which that took quote long also. LapLink was a great tool for that, before LAN became more common.
2 comments
WarOnPrivacy 14 days ago
I had Spinrite III; Gibson made good stuff. But I was often in situations where all I had was a WD, etc MFM controller.
link
mrlonglong 15 days ago
Sigh. How the times change. Right now I'm looking at 2 days to rebuild my NAS array after adding another 10TB disk.
link
mesrik 14 days ago
True, that's some things have changed.

But it's when you consider taking first full drive backup times. It still takes 'forever' when you'd rather like to get on with fixing single disk issues. Or like when you wait RAID rebuilds too.

We did plan our maintenance tasks so, that there was something else useful we could do while obligatory backup and Spinrite was running. Upgraded another devices EPROM's, PC BIOS or dot matrix printer EPROM, cleaned printer heads, chaff and dust from chassis, changed colour ribbon to it. Ran updates to software we had sold etc. And went lunch with the customer:)

link
mrlonglong 12 days ago
It turned into 5 days. After realising I'd forgotten to change the block size to 4096 and it was horribly slow, I yanked the drive out, tried with my utility to set it to 4096 bytes and failed, found another utility that did the job, put the drive back in, wait 3 days for it to rebuild the array without the new disk, and then another two days to add the disk back in. Sigh.

Still, with 4k block sizes, it goes like shit off a shovel.

link